tag:blogger.com,1999:blog-2763891945076387742023-11-16T08:52:20.054+01:00StatconStatistik-Software, Training & BeratungKulighttp://www.blogger.com/profile/08848878437446433017noreply@blogger.comBlogger28125tag:blogger.com,1999:blog-276389194507638774.post-79005900735986426092015-06-16T12:56:00.000+02:002015-06-16T13:10:46.543+02:00How to Choose the Right Screening Design!<div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiFKnjX5j76DoUen7Ny5b-x3t3PtKmbtYjD0cEFVuzruxTVgSA-ua1cm7uimiadRCNRqyZRI-yR_6d-wj0gRjtbIMBP05tMZsqlEPXIxyxOabDiMeeQuLvxAmTad3fAUftX8XJa-grkRn8/s1600/Header.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="228" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiFKnjX5j76DoUen7Ny5b-x3t3PtKmbtYjD0cEFVuzruxTVgSA-ua1cm7uimiadRCNRqyZRI-yR_6d-wj0gRjtbIMBP05tMZsqlEPXIxyxOabDiMeeQuLvxAmTad3fAUftX8XJa-grkRn8/s400/Header.png" width="400" /></a></div><div class="separator" style="clear: both; text-align: center;"><br />
</div>New DoE-users often get overwhelmed by the amount of different designs that are available. Instead of seeing the benefits of just using the DoE (Design of Experiments) approach at all the tend to worry how to choose the right design for a given application. In this piece I will present my approach to choose the right screening design while introducing a JMP-script that supports you with that.<br />
<br />
<a name='more'></a>When setting up a design to make some research about an unknown process we often start with so called screening designs. These are designs allowing you to analyze the effect of many factors with a relatively small number of experiments. Over the years I developed a small checklist that helps me keeping track of the important decision-points when setting up a screening design.<br />
<br />
The very first subsection of my checklist are reconsiderations:<br />
<br />
<ol><li>What is my overall goal?</li>
<li>What are my responses? How do I measure them?</li>
<li>What factors should be considered? In which ranges?</li>
<li>Are there block-factors?</li>
<li>Is it a split-plot problem?</li>
<li>Are there problematic combinations of factor settings?</li>
</ol><div>Let's assume we are doing <a href="http://www.paperhelicopterexperiment.com/" target="_blank">the paper helicopter experiment</a>. </div><div><ol><li><b>Goal:</b> Maximize the flight time.</li>
<li><b>Response:</b> Flight time (s/10)</li>
<li>Of course the <b>factors</b> that are chosen vary from time to time. The factors in table 1 are the ones I used in a training in Berlin last week.</li>
<li><b>No block-factors.</b></li>
<li><b>No split-plots.</b></li>
<li><b>No problematic combinations of factor settings.</b></li>
</ol></div><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgoQ2il8S0YC2GXQ9xKRDujngTe9f0_aROp0VvJliJ4j6dzPQCej2-DF0UJYlycfCuGl30K2ilSG9z5kUkaEfA2ltQzaMa5towGlObbYwHaqxalPDHfGRCB698RZXX4x4YPVdd4MxMTjKs/s1600/factors.PNG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgoQ2il8S0YC2GXQ9xKRDujngTe9f0_aROp0VvJliJ4j6dzPQCej2-DF0UJYlycfCuGl30K2ilSG9z5kUkaEfA2ltQzaMa5towGlObbYwHaqxalPDHfGRCB698RZXX4x4YPVdd4MxMTjKs/s1600/factors.PNG" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Table 1: Factors for the Paper Helicopter Experiment</td></tr>
</tbody></table><div class="separator" style="clear: both; text-align: left;">All of those reconsiderations are very crucial to the success of a DoE - and none of them requires statistics knowledge. This is why I - as a statistician - want to focus on the next step: </div><div class="separator" style="clear: both; text-align: left;"><br />
</div><b><u>Design Choice</u></b><br />
<div>After we did all the hard work, the rest should be easy. We just need to figure out which design is the right one for our problem. Of course there are many different screening designs. Depending on which software you use you will see:</div><div><ul><li><a href="https://en.wikipedia.org/wiki/Factorial_experiment" target="_blank">Full-Factorial Designs</a></li>
<li><a href="https://en.wikipedia.org/wiki/Fractional_factorial_design" target="_blank">Fractional-Factorials</a></li>
<li><a href="http://www.jmp.com/en_gb/whitepapers/three-level-definitive-screening.html" target="_blank">Definitive Screenings</a></li>
<li><a href="https://en.wikipedia.org/wiki/Plackett%E2%80%93Burman_design" target="_blank">Plackett-Burman</a></li>
<li><a href="https://en.wikipedia.org/wiki/Orthogonal_array" target="_blank">Orthogonal Arrays</a></li>
<li><a href="https://en.wikipedia.org/wiki/Optimal_design" target="_blank">D-Optimal Designs</a></li>
<li><a href="http://mnasq.org/wp-content/uploads/DOE-Developments-Feb-2015-MN-ASQ.pdf" target="_blank">Min-Run Resolution V, IV</a></li>
<li>...</li>
</ul><div>To decide for one of those it is of course useful to understand the specific designs and nothing is better than knowing the strengths and weaknesses of each individual design. Non the less I want to provide a more general approach to chose a design even without knowing each individual design. Therefore I will compare all of these designs in terms of <b>power</b>, <b>aliasing</b> and <b>size</b>. In most DoE-software packages this involves setting up each individual design to do power-analysis and get an aliasing-report. To ease that step I created a script that will do most of the work for you in JMP (You will find the script <a href="https://community.jmp.com/docs/DOC-7501" target="_blank">here</a>).</div></div><div><br />
</div><div><u><b>Using the JMP-Script</b></u></div><div>The script requires you to set up two data tables before we may start. The first table should contain all the information about your responses.</div><div><br />
</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEidKUHg3d8q4G4VneWQMD-yIqqCGe2g6GT6ZO5Z2V1OcFpRI0cEAjG96iqKK7doAzc2bL5plsAr1YaHvc6SvVkVGZIb4NqjBZ6ZKmn_Ce7KBNwbBUSkDLCG8QQBIr9Bj7G58tzzobx3t_s/s1600/responses+data+table.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="278" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEidKUHg3d8q4G4VneWQMD-yIqqCGe2g6GT6ZO5Z2V1OcFpRI0cEAjG96iqKK7doAzc2bL5plsAr1YaHvc6SvVkVGZIb4NqjBZ6ZKmn_Ce7KBNwbBUSkDLCG8QQBIr9Bj7G58tzzobx3t_s/s640/responses+data+table.PNG" width="640" /></a></div><div><br />
</div><div>The second table has a different structure and cares for the factors of your problem.</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjZF6-EPM54OrjkuYLbFAyPEBx4G9I3HG_OQNZcS_tGn60g_vfypEY-pSgH3975Sh35RwrWv3Bw_O-07f-yHftsXmVlf5w4IsXm9Q2njkXa85SzU_sOsJig23BpbHffoYXqCojmTSf_MDo/s1600/factors+data+table.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="378" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjZF6-EPM54OrjkuYLbFAyPEBx4G9I3HG_OQNZcS_tGn60g_vfypEY-pSgH3975Sh35RwrWv3Bw_O-07f-yHftsXmVlf5w4IsXm9Q2njkXa85SzU_sOsJig23BpbHffoYXqCojmTSf_MDo/s640/factors+data+table.PNG" width="640" /></a></div><div><br />
</div><div>Probably the easiest way to create these tables is to go to the <b><i>Main Menu -> DOE -> Custom Designer</i></b></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkcPeuuyDvYcMqYOCUsHTAkC2KWhmzmtbtuMC809C39VgeEYwnq1k8MKgbb8wTKEuDgnJz044tsaeLVEmjZnNEvSjkO4RFzKpxab_9TXsRGb38PEum7IKok13q9U9ytbdpigcwT8QMOs4/s1600/custom+designer.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="462" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkcPeuuyDvYcMqYOCUsHTAkC2KWhmzmtbtuMC809C39VgeEYwnq1k8MKgbb8wTKEuDgnJz044tsaeLVEmjZnNEvSjkO4RFzKpxab_9TXsRGb38PEum7IKok13q9U9ytbdpigcwT8QMOs4/s640/custom+designer.PNG" width="640" /></a></div><div><br />
</div><div>In there enter all responses and factors as you are used to it. Then use the hotspot at <i><b>Custom Design </b></i>(the red triangle) and <b><i>Save Factors</i></b> and <i><b>Save Responses</b></i>.</div><div><br />
</div><div>Now we are ready to go. Open the script, right click somewhere in the script and klick on <b><i>Run Script</i></b>.</div><div><br />
</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEinZufG-NG4Wg28m6fagz7ZmHnqAC6qy_S0cMaQChVvPaIKunWZ6R0N0GrpXL9st04Vt50IVa5S1gcv72Y37ZgsjDfCqLJxQg5Hjy7Up1DrgWKmrbrslUhnMPxprktAzgAVfWrtypFgsrk/s1600/dialog.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEinZufG-NG4Wg28m6fagz7ZmHnqAC6qy_S0cMaQChVvPaIKunWZ6R0N0GrpXL9st04Vt50IVa5S1gcv72Y37ZgsjDfCqLJxQg5Hjy7Up1DrgWKmrbrslUhnMPxprktAzgAVfWrtypFgsrk/s1600/dialog.PNG" /></a></div><div>Now we have to provide some additional information for the power analysis. </div><div><ul><li>What is the desired level of significance?</li>
<li>What is you anticipated coefficient (basically: relevant difference/2)?</li>
<li>And what is your experimental error?</li>
</ul><div>For the helicopter experiment a <b>difference to detect of 0.5s </b>(thus a <b>coefficient of 0.25s</b>) sounds reasonable for me. Having done this experiment probably 50 times already I have some amount of experience with the<b> experimental error</b>. It tends to be around 1/3rd of a second. Usually I plan a bit conservatively: <b>0.35s</b> might be a good value. If you are not completely familiar with power analysis <a href="http://blogs.sas.com/content/jmp/2012/04/02/fundamentals-of-power-analysis-in-experiment-design/" target="_blank">this</a> and <a href="http://blogs.sas.com/content/jmp/2012/04/09/revised-in-jmp-10-power-analysis-in-custom-design/" target="_blank">this</a> might be a good starting points for you to read.</div></div><div><br />
</div><div>The last thing we have to do, is to tell JMP which data set contains the information about the responses and which one is about the factors. In my case <b><i>Untitled 2 </i></b>was for the response, <b><i>Untitled</i></b> was for the factors. The script isn't actually the fastest piece of code - apologies for that. After a while you should see this report:</div><div><br />
</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhe2nKzrl_D0cl63smnGJV00oX94lIkyq_0SyvqBD38UrkF_brGPdoyWo6qPfQdMqmb9RVkRA8zGVPC-YgGL8ANYSyBYPiSDAvmBBZ44OEM_mGF46JRVSHqZJCCGQq8J3QgI02Oo9VuM_s/s1600/report.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="302" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhe2nKzrl_D0cl63smnGJV00oX94lIkyq_0SyvqBD38UrkF_brGPdoyWo6qPfQdMqmb9RVkRA8zGVPC-YgGL8ANYSyBYPiSDAvmBBZ44OEM_mGF46JRVSHqZJCCGQq8J3QgI02Oo9VuM_s/s640/report.PNG" width="640" /></a></div><div><br />
</div><div>The first section gives you the power for your different designs. All values below 0.8 are red, all other values are green. Usually we want screening experiments to achieve a power of at least 80%. For our example we see that the <b>Definitive Screening</b> and the <b>Fractional Factorial Resolution 3</b> are out of contention. </div><div><br />
</div><div>You might want to check if the aliasing of the Fractional Factorial Resolution 4 is ok for your purposes. Therefore open the corresponding outline item.</div><div><br />
</div><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg5kKNOQe8Y7MnsbWvVGvsSeDtkPldyr63aoyC0lJu2pODsCWcIkZQHfMpk0Sfk9BC0pdnwSp0QwASff7SzO2B_GdGeAvKqnIjU67Fx1hwXj3dURdV8N_ASk5j15eWqZXVJWLvgMuvtyqM/s1600/frac4+aliasing.PNG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg5kKNOQe8Y7MnsbWvVGvsSeDtkPldyr63aoyC0lJu2pODsCWcIkZQHfMpk0Sfk9BC0pdnwSp0QwASff7SzO2B_GdGeAvKqnIjU67Fx1hwXj3dURdV8N_ASk5j15eWqZXVJWLvgMuvtyqM/s1600/frac4+aliasing.PNG" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Colormap of Correlations - Fractional Factorial Resolution 4</td></tr>
</tbody></table>As you can see there is quite some aliasing going on in this design. Personally I would not feel comfortable having some aliasing between 2-factor-interactions like <i>Wing Length*Wing Width </i>(=airofoil) or <i>Wing Length*Weight</i>. Thus the Fractional Factorial Design Resolution 4 does not seem to be an option, neither.<br />
<div><br />
</div><div>As the <b>Full Factorial</b> design, like the <b>Fractional Factorial Design Resolution 5</b> have no issues with aliasing at all, I recommend using the<b> Fractional Factorial Res. 5 </b>in this case. It uses less runs but still has sufficient power.</div><div><br />
</div><div>To get the design just press the button Fractional Factorial (Res.5) at the bottom of the report and it will appear. </div><div><br />
</div><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjocXgaCTCu9kdw-EG6FUyqrlFW_NJIojJbb6K2lpBHhwJr-Gxh7qqaN7i4YChockw7BhjS2BuvQ3Yk-UBfu1mkI5aMkRBpNWYzDtKTTrHAHjGJP1PQbIFe9oFIE6SL8z3ARmaKu4Y2M8Y/s1600/frac5+design.PNG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="424" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjocXgaCTCu9kdw-EG6FUyqrlFW_NJIojJbb6K2lpBHhwJr-Gxh7qqaN7i4YChockw7BhjS2BuvQ3Yk-UBfu1mkI5aMkRBpNWYzDtKTTrHAHjGJP1PQbIFe9oFIE6SL8z3ARmaKu4Y2M8Y/s640/frac5+design.PNG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Fractional Factorial Design Resolution 5</td></tr>
</tbody></table><div>Time for some helicopter producing! Have fun!</div><div><br />
</div><div>P.S.: Currently the script only supports these five different designs. But it is easy to add other design. Just tell me about your preferred screening designs in JMP and I will add it whenever I have the time!</div><div></div>Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com2tag:blogger.com,1999:blog-276389194507638774.post-71902477301157435912015-05-04T13:39:00.002+02:002015-05-05T13:55:56.276+02:00Exploring Spatial Autocorrelation: Moran's I and Geary's Ratio<span style="font-family: Verdana, sans-serif;">While using autocorrelation statistics for time series data is quite common, one has to dig a bit deeper to evaluate spatial autocorrelation for some given data. Spatial autocorrelation might be a starting-point for any analysis of spatial data to get a first impression if places that are close to each other are similar in regards of a variable of interest.</span><br />
<span style="font-family: Verdana, sans-serif;"><br />
</span> <span style="font-family: Verdana, sans-serif;">This piece shows how to use the <a href="https://community.jmp.com/docs/DOC-7331" target="_blank">Spatial Data Analysis-Add-In</a> (Version 0.92) for JMP to calculate the most commonly used metrics to measure spatial autocorrelation: <a href="http://en.wikipedia.org/wiki/Moran's_I" target="_blank">Moran's I</a> and <a href="http://en.wikipedia.org/wiki/Geary's_C" target="_blank">Geary's Ratio</a>.</span><br />
<span style="font-family: Verdana, sans-serif;"><br />
</span> <br />
<div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjA_ZF48IlY59gh0fZWZc5uMNWYNHKlUZdS7Tep3rNTDd6fY2IRtwXNt6u1gB1EeS6a0SXJpbdfTyKqD1NbAzQPMDWgH6PvHn7BgLP2N1KUtHY0LjqPHLxoVktToNuT6Kbds_xVDebm75M/s1600/LA+Map+small.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><span style="font-family: Verdana, sans-serif;"><img border="0" height="300" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjA_ZF48IlY59gh0fZWZc5uMNWYNHKlUZdS7Tep3rNTDd6fY2IRtwXNt6u1gB1EeS6a0SXJpbdfTyKqD1NbAzQPMDWgH6PvHn7BgLP2N1KUtHY0LjqPHLxoVktToNuT6Kbds_xVDebm75M/s640/LA+Map+small.png" width="640" /></span></a></div>
<span style="font-family: Verdana, sans-serif;"><br />
</span> <br />
<a name='more'></a></div>
<div>
<span style="font-family: Verdana, sans-serif;">Like <a href="http://www.ats.ucla.edu/stat/r/faq/morans_i.htm" target="_blank">here</a> we will try to figure out if ozone measurements at 32 different locations in Los Angeles are spatially correlated or not.</span><br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi4F1YDkiUpJMWJ6ya9sUZv9okopi5Z22Nv5gOy0VY0OjQgZ4N6qRYhg9RKQUshcHn0wg4vVVqLfe_rrm0fmvo2rrhCEKdkMSzvKY0QQHQqqbsSh0ett1yJ8p8fUDIJvNFGs9IOdnJ9gXQ/s1600/Data.PNG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><span style="font-family: Verdana, sans-serif;"><img border="0" height="594" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi4F1YDkiUpJMWJ6ya9sUZv9okopi5Z22Nv5gOy0VY0OjQgZ4N6qRYhg9RKQUshcHn0wg4vVVqLfe_rrm0fmvo2rrhCEKdkMSzvKY0QQHQqqbsSh0ett1yJ8p8fUDIJvNFGs9IOdnJ9gXQ/s640/Data.PNG" width="640" /></span></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-family: Verdana, sans-serif;">Ozone Data Set</span></td></tr>
</tbody></table>
<span style="font-family: Verdana, sans-serif;">To calculate Moran's I and Geary's Ratio we first need to create a weight-matrix. This weight-matrix shall contain values representing the spatial similarity between data points. Thus the cell in the first row and second column will represent how close data points one and two are. Here we will just calculate the euclidean distances and use the inverse as a measure of similarity. </span><span style="font-family: Verdana, sans-serif;">Typically we will set the similarity of a datapoint with itself to zero. Thus the diagonal of this matrix is all zeros.</span><br />
<div>
<br /></div>
<span style="font-family: Verdana, sans-serif;">$$w_{i,j} = \begin{cases} i \neq j & \frac{1}{\sqrt{(x_i-x_j)^2 + (y_i-y_j)^2}} \\ i=j & 0\end{cases}$$</span><br />
<span style="font-family: Verdana, sans-serif;"><br /></span>
<span style="font-family: Verdana, sans-serif;">Last but not least we will typically row standardize the matrix - meaning: Divide all rows by the sum of all values in that row. This makes sure the overall sum of each row is equal to 1.</span><br />
<span style="font-family: Verdana, sans-serif;"><br />
</span> <span style="font-family: Verdana, sans-serif;">To ease this process I added a new feature to the addin. Go to <b><i>Add-Ins => Spatial Data Analysis => Euclidean Distance Matrix</i></b>. As a result you will get this prompt:</span><br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<span style="font-family: Verdana, sans-serif; margin-left: 1em; margin-right: 1em;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRpY-GkOj-bH3aURUKIix131JETWzGF9cYYocftevIVnUQcZOmkV3JOs-UWd4AYeFXWti8dper9jHvsqZIqKavgrgqa_HKmL9pQvu0I3t9xJgsW0K38SijayQ3t73az0l9oZCTAl0gBug/s1600/dialog+dist+matrix.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="160" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjRpY-GkOj-bH3aURUKIix131JETWzGF9cYYocftevIVnUQcZOmkV3JOs-UWd4AYeFXWti8dper9jHvsqZIqKavgrgqa_HKmL9pQvu0I3t9xJgsW0K38SijayQ3t73az0l9oZCTAl0gBug/s400/dialog+dist+matrix.PNG" width="400" /></a></span></div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<span style="font-family: Verdana, sans-serif;">Usually it will just create a distance matrix for all rows in the data table, using euclidean distances. But selecting the checkbox <i><b>Similarity Weights</b> </i>will calculate the weights like above. Finally we receive a dataset in JMP like the following:</span><br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhSusTFpc99um0725h0wJXbIOsLEHmpIwKsV4BZkyklMM2d676l9fEuTPYdGuHYCtPt2UmZumbJ8mEVwvuvmnIUIhXRv5svdbNgYtQ1xRk7sMj5iPOGTpk-YIpKN3tnt3wu92YNQrJfQoU/s1600/dist+matrix.PNG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><span style="font-family: Verdana, sans-serif;"><img border="0" height="452" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhSusTFpc99um0725h0wJXbIOsLEHmpIwKsV4BZkyklMM2d676l9fEuTPYdGuHYCtPt2UmZumbJ8mEVwvuvmnIUIhXRv5svdbNgYtQ1xRk7sMj5iPOGTpk-YIpKN3tnt3wu92YNQrJfQoU/s640/dist+matrix.PNG" width="640" /></span></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-family: Verdana, sans-serif;">Distance Matrix for Ozone Data</span></td></tr>
</tbody></table>
<span style="font-family: Verdana, sans-serif;">Of course this is not the only way to measure the spatial <i>closeness</i> of data points. A future blogpost might address this topic and introduce additional features to generate different kind of weight matrices using the JMP Add-In.</span><br />
<br />
<h4>
<span style="font-family: Verdana, sans-serif;">Moran's I and Geary's Ratio</span></h4>
</div>
<div>
<span style="font-family: Verdana, sans-serif;">Now that we have the similarities we might start to think about the coefficients of spatial autocorrelation. Without going into the details, here is their interpretation:</span><br />
<b style="font-family: Verdana, sans-serif;"><br /></b>
<b style="font-family: Verdana, sans-serif;">Moran's I</b></div>
<div>
<ul>
<li><span style="font-family: Verdana, sans-serif;">Moran's I is between</span><span style="font-family: Verdana, sans-serif;"> -1 and 1</span><span style="font-family: Verdana, sans-serif;"> (as long as your weight matrix is row-standardized). </span></li>
<li><span style="font-family: Verdana, sans-serif;">Values </span><span style="font-family: Verdana, sans-serif;">close to 0 indicate no spatial autocorrelation*</span><span style="font-family: Verdana, sans-serif;">. </span></li>
<li><span style="font-family: Verdana, sans-serif;">Values </span><span style="font-family: Verdana, sans-serif;">close to 1 indicate strong positive spatial autocorrelation</span><span style="font-family: Verdana, sans-serif;">. I.e. regions close to each other behave similar in terms of the variable of interest. </span></li>
<li><span style="font-family: Verdana, sans-serif;">Values </span><span style="font-family: Verdana, sans-serif;">close to -1 indicate strong negative spatial autocorrelation</span><span style="font-family: Verdana, sans-serif;">.</span></li>
</ul>
<b style="font-family: Verdana, sans-serif;"></b><br />
<div>
<b style="font-family: Verdana, sans-serif;"><i style="font-family: Times; font-weight: normal;"><span style="font-family: Verdana, sans-serif;">* Actually the expected value of Moran's I if there is no spatial autocorrelation is not exactly 0. The expected value if there is no spatial autocorrelation is $E(I) = \frac{-1}{N-1}$, which is typically close to 0.</span></i></b></div>
<b style="font-family: Verdana, sans-serif;">
<div>
<i style="font-family: Times; font-weight: normal;"><span style="font-family: Verdana, sans-serif;"><br /></span></i></div>
</b><b style="font-family: Verdana, sans-serif;">Geary's Ratio</b></div>
<div>
<ul>
<li><span style="font-family: Verdana, sans-serif;">Geary's ratio is between </span><span style="font-family: Verdana, sans-serif;">0 and 2</span><span style="font-family: Verdana, sans-serif;">. </span></li>
<li><span style="font-family: Verdana, sans-serif;">Values </span><span style="font-family: Verdana, sans-serif;">close to 1 indicate no spatial</span><span style="font-family: Verdana, sans-serif;"> </span><span style="font-family: Verdana, sans-serif;">autocorrelation</span><span style="font-family: Verdana, sans-serif;">. </span></li>
<li><span style="font-family: Verdana, sans-serif;">Values</span><span style="font-family: Verdana, sans-serif;"> close to 0 indicate strong positive autocorrelation.</span></li>
<li><span style="font-family: Verdana, sans-serif;">Values </span><span style="font-family: Verdana, sans-serif;">close to 2 indicate strong negative spatial autocorrelation</span><span style="font-family: Verdana, sans-serif;">.</span></li>
</ul>
</div>
<div>
<span style="font-family: Verdana, sans-serif;">To get those numbers for the given data go to </span><b style="font-family: Verdana, sans-serif;"><i>Add-Ins => Spatial Data Analysis => Spatial Autocorrelation</i></b><span style="font-family: Verdana, sans-serif;">.</span></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjsVZp-KTzidP0-HBCgx681HHQMsv8zsQGvZrHXqN-oETlx9HsJ6PG2dbQCfVmP3ruSW3e9gMirKTrB_XPsChSr_gc9KSiln5vGXkyR168zR-8igRR9d60cYCfs4_dX5a_kmY108Vy4kp4/s1600/sac+dialog.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><span style="font-family: Verdana, sans-serif;"><img border="0" height="267" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjsVZp-KTzidP0-HBCgx681HHQMsv8zsQGvZrHXqN-oETlx9HsJ6PG2dbQCfVmP3ruSW3e9gMirKTrB_XPsChSr_gc9KSiln5vGXkyR168zR-8igRR9d60cYCfs4_dX5a_kmY108Vy4kp4/s400/sac+dialog.PNG" width="400" /></span></a></div>
<div>
<span style="font-family: Verdana, sans-serif;"><br />
</span></div>
<div>
<span style="font-family: Verdana, sans-serif;">The dialog asks for:</span></div>
<div>
<ul>
<li><span style="font-family: Verdana, sans-serif;">The <b>variable</b> of interest. For now these are the ozone measurements stored in column <i><b>Av8top</b></i>.</span></li>
<li><span style="font-family: Verdana, sans-serif;">Which <b>measures</b> to calculate: <b>Moran's I</b> and/or <b>Geary's Ratio</b>.</span></li>
<li><span style="font-family: Verdana, sans-serif;">The previously calculated <b>weight matrix</b> as a JMP-data-table.</span></li>
</ul>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgjsWNDK8RyxfhJoldc5tCVJTp7zoXstl_zJxGxjTM9FZhTPJleXuJpZgA-ZbZN2WIdVEwf-JO8yOcWSgGnR9x6cGRLadUG8RLFsym2ZohQn7cmnXhlxe6402n7BhlFwqcTqCqCE7Njqno/s1600/report.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><span style="font-family: Verdana, sans-serif;"><img border="0" height="380" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgjsWNDK8RyxfhJoldc5tCVJTp7zoXstl_zJxGxjTM9FZhTPJleXuJpZgA-ZbZN2WIdVEwf-JO8yOcWSgGnR9x6cGRLadUG8RLFsym2ZohQn7cmnXhlxe6402n7BhlFwqcTqCqCE7Njqno/s640/report.PNG" width="640" /></span></a></div>
<div>
<span style="font-family: Verdana, sans-serif;"><br />
</span></div>
</div>
<div>
<span style="font-family: Verdana, sans-serif;">The report shows that the ozone-measurments are somewhat positively correlated. Being at roughly <b>0.23</b> Moran's I is (significantly) larger than -0.0323, which would be the expected value of Moran's I if there was no spatial autocorrelation.</span></div>
<div>
<span style="font-family: Verdana, sans-serif;"><br />
</span></div>
<div>
<span style="font-family: Verdana, sans-serif;">With a value of <b>0.77 </b>Geary's Ratio is (significantly) smaller than 1. This indicates positive spatial autocorrelation.</span></div>
<div>
<span style="font-family: Verdana, sans-serif;"><br />
</span></div>
<div>
<span style="font-family: Verdana, sans-serif;">Don't become confused by the three color maps at the bottom part of the dialog. They are not depending on your data. Their only purpose is to give people an idea how to interpret Moran's I and Geary's Ratio by showing how spatially correlated data looks like. </span></div>
<div>
<ul>
<li><span style="font-family: Verdana, sans-serif;"><b><u>Positive Spatial Correlation (left hand side):</u></b> The data is clustered in terms of the variable of interest.</span></li>
<li><span style="font-family: Verdana, sans-serif;"><b><u>No Spatial Correlation (middle): </u></b>There is no spatial structure in the data.</span></li>
<li><span style="font-family: Verdana, sans-serif;"><b><u>Negative Spatial Correlation (right hand side):</u></b> Regions that are close to each other tend to show different values for the variable of interest.</span></li>
</ul>
<div>
<span style="font-family: Verdana, sans-serif;"><br />
</span></div>
</div>
<div>
<br /></div>
Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-64931213699469973312015-04-15T16:04:00.001+02:002015-05-04T08:44:10.918+02:00Solving Traveling Salesperson Problems with JMPThis is part 2 of my series about spatial data analysis with JMP. After we learned how to geocode addresses in <a href="http://statistiksoftware.blogspot.de/2015/03/spatial-data-analysis-with-jmp.html" target="_blank">part 1</a>. I will now show you how to solve <b>T</b>raveling <b>S</b>alesperson <b>P</b>roblems using the JMP Add-In.<br />
<br />
<a name='more'></a>Remember the previous example? When we where at Brussels for the <a href="https://www.jmp.com/about/events/summit-europe/live.shtml" target="_blank">JMP Discovery Summit</a> it was for sure nice to see all the sights in brussels on a map in JMP. But wouldn't it be even better to see the most efficient order in which we should visit them? Thats what TSPs - or Traveling Salesperson Problems - are all about. If you want to visit k cities: What is the fastest way to do that? While the problem might sound simple it is fairly computerintensive to solve.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg1xb7r2X1yDkI7WY7cjrbtADirkwUxgjQHeZsYXJZYiYr_ouHQ8Rk06jSN1PzPd993yyvAWL4ve39Pitmdu5mJWECECfFVfaoYvrId99TYufAWkP3r3aMplUooYvh2Qlo2Pc3LoScyuVo/s1600/map2.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg1xb7r2X1yDkI7WY7cjrbtADirkwUxgjQHeZsYXJZYiYr_ouHQ8Rk06jSN1PzPd993yyvAWL4ve39Pitmdu5mJWECECfFVfaoYvrId99TYufAWkP3r3aMplUooYvh2Qlo2Pc3LoScyuVo/s1600/map2.PNG" height="608" width="640" /></a></div>
<br />
<br />
Similarly like for the geocoding I was able to find some R-functions that do the hard work for us (r-package: TSP from Hahsler and Hornik). Thus for us it's just a matter of a few clicks and time!<br />
<br />
For the example I will use a smaller dataset though. First of all it is not realistic to visit all 42 places in a day and: <span style="color: red;">Using the whole data you will need to get 42*42 distances. These are 1764 queries for the </span><b style="color: red;">google maps API</b><span style="color: red;">. If you are not a commercial user this will need most of the 2500 free queries that you get every day!</span><br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<br />
Let's focus on our 5 most favorite comics:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgCeex-Z3CKxy0MPQeU5FEuhwfc34TXdBAc52XRQxi-H06jy8_vkif8NYxJo3hzXfo7KsVlbSw9kJJEwixVFMQwY48K4I44gd29EPA_G13TWbYhEXXZnWYTmWVxZKnrCKm1jh5ijg102V8/s1600/data.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgCeex-Z3CKxy0MPQeU5FEuhwfc34TXdBAc52XRQxi-H06jy8_vkif8NYxJo3hzXfo7KsVlbSw9kJJEwixVFMQwY48K4I44gd29EPA_G13TWbYhEXXZnWYTmWVxZKnrCKm1jh5ijg102V8/s1600/data.PNG" height="266" width="640" /></a></div>
<br />
Remember: Column Address contains all the addresses from the Brussels Comic Tour and that is all we need. Just select <b><i>Add-Ins -> Spatial Data Analysis -> TSP</i></b> in JMP's main menu.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhR8Qlnv6MZTb_RiDi94Hi67CUaDyVIDkWTVN0hegN37uHprFqEdJXiESc0BGN2aElyGWxzHx8xye3hS7vp4bknbIPsb-aqx8vFUicCYGi1pdKf9GNhdPW3I9UL44X3EupW5HBN9Mgi-LU/s1600/tsp+dialog.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhR8Qlnv6MZTb_RiDi94Hi67CUaDyVIDkWTVN0hegN37uHprFqEdJXiESc0BGN2aElyGWxzHx8xye3hS7vp4bknbIPsb-aqx8vFUicCYGi1pdKf9GNhdPW3I9UL44X3EupW5HBN9Mgi-LU/s1600/tsp+dialog.PNG" height="502" width="640" /></a></div>
<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
The dialog will ask you for the column containing the addresses for all places you want to visit. By the way: There is no need to do the geocoding first. So that is easy, but what is behind the advanced settings?<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgj0kxZTVy_RWQNl9kMmCU3htA-hyCHIyu-6AZUh4T8T2FxExmuEHiKRc-hu2CI2sENA5Dl1bwhIPkl3flFhhPfqo8TLYSbU1hv0T-3tMrxBWjC6gKh0ZYZlpGuEjTKvyEAsX3a8c717QY/s1600/tsp+dialog2.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgj0kxZTVy_RWQNl9kMmCU3htA-hyCHIyu-6AZUh4T8T2FxExmuEHiKRc-hu2CI2sENA5Dl1bwhIPkl3flFhhPfqo8TLYSbU1hv0T-3tMrxBWjC6gKh0ZYZlpGuEjTKvyEAsX3a8c717QY/s1600/tsp+dialog2.PNG" height="500" width="640" /></a></div>
<br />
There are 3 sections in the advanced settings:<br />
<ol>
<li><b>Metrics</b>: Are you going to walk by foot, use a bike or the car? Do you want to optimize (minimize) the distance (meters, miles) or the time (minutes) it takes to visit all places?</li>
<li><b>Algorithms</b>: This is more technical and I would like to refer to the <a href="http://cran.r-project.org/web/packages/TSP/vignettes/TSP.pdf" target="_blank">documentation of the TSP-package</a> if you are interested in the details of the different algorithms. In my experience it is often useful to give multiple algorithms a try, as they might give you different results for your problem.</li>
<li>The last section asks if JMP should <b>visualize the final route</b>. Just give it a try and select the checkbox.<br />The <b>number of iterations</b> is controlling how often the used algorithms is used. This is needed as the algorithms do not always find the best solution on the first try. For our 5 places it shouldn't be that hard to calculate it a couple of times. Let's use the default 25.</li>
</ol>
<div>
After a couple of seconds you will get this result:</div>
<div>
<br /></div>
<div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgfCXAcUTcEaMTV8PvI0L5yNCwneFPGmx__PUvllnkuPU9tsgtGUPvVRZSiMlI1zI-gM6gZ_r2DkTVc6iPLIGI7ZQPExMjXk-GITS_k4XokGHaKh97TmuUwD_Pe54G0MZfJxiAG5MfTOx4/s1600/result.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgfCXAcUTcEaMTV8PvI0L5yNCwneFPGmx__PUvllnkuPU9tsgtGUPvVRZSiMlI1zI-gM6gZ_r2DkTVc6iPLIGI7ZQPExMjXk-GITS_k4XokGHaKh97TmuUwD_Pe54G0MZfJxiAG5MfTOx4/s1600/result.PNG" /></a></div>
<br />
Now the main data set is sorted in the optimal order.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjf5idz3Nci_gMwh07lsRIIhyatDi2G5zpPv-oAQjZLsuIZz1NF-NAjnm5rWodHbQrvWhopuA9DtRkRdg_3XQtzdPeCjb137OuLk4NMsBg4pcTPjoMCrihxYeHSjVY4WjoKcMIKVeiL9DE/s1600/sorted.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjf5idz3Nci_gMwh07lsRIIhyatDi2G5zpPv-oAQjZLsuIZz1NF-NAjnm5rWodHbQrvWhopuA9DtRkRdg_3XQtzdPeCjb137OuLk4NMsBg4pcTPjoMCrihxYeHSjVY4WjoKcMIKVeiL9DE/s1600/sorted.PNG" height="430" width="640" /></a></div>
<br /></div>
And it's easy to visualize this solution in the graph builder, using points and lines (in row order) - of course after geocoding the addresses first:<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrIG0_D1ABfus8gcsVQC4eI8Arv5PPyalL11Xc2AZTEkWBZF5SfZoCTFxAa2YFPbYLE-WvFesqbIXSyyba3iWA8xOEsxN-hN18bJNzD1WcUADt1E28hzfDPSTxYm0e9vxr1PK2GbJwZjA/s1600/simple+graph.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrIG0_D1ABfus8gcsVQC4eI8Arv5PPyalL11Xc2AZTEkWBZF5SfZoCTFxAa2YFPbYLE-WvFesqbIXSyyba3iWA8xOEsxN-hN18bJNzD1WcUADt1E28hzfDPSTxYm0e9vxr1PK2GbJwZjA/s1600/simple+graph.PNG" height="556" width="640" /></a></div>
<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<br />
If you pressed the Display Route-checkbox, you will get this more detailed graph of the route automatically:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhDKvuV5ImWEalz-vbv3XB1Idcw3hS-n_Oc1LXb-a1eafEdthT3_ur_28TvbC4ZF1q3LA852Fix9wNkQQEaUKUiQwY7LBnTHjHuneD9enyON-OVo0a8dQ8NZAXWFib0s1s0yOWXCRGodu8/s1600/better+graph.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhDKvuV5ImWEalz-vbv3XB1Idcw3hS-n_Oc1LXb-a1eafEdthT3_ur_28TvbC4ZF1q3LA852Fix9wNkQQEaUKUiQwY7LBnTHjHuneD9enyON-OVo0a8dQ8NZAXWFib0s1s0yOWXCRGodu8/s1600/better+graph.PNG" height="636" width="640" /></a></div>
<br />
<b>Literature</b><br />
<br />
<ol>
<li><b>Michael Hahsler and Kurt Hornik (2015).</b><i> TSP: Traveling Salesperson Problem (TSP)</i>. R package version 1.0-10. http://CRAN.R-project.org/package=TSP</li>
<li><b>Michael Hahsler, and Kurt Hornik (2007)</b>, <i>TSP - Infrastructure for the traveling salesperson problem</i>. Journal of Statistical Software 23/2. URL: http://www.jstatsoft.org/v23/i02/.</li>
</ol>
Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-46964055469010422312015-03-27T14:26:00.003+01:002015-03-27T14:27:44.973+01:00Spatial Data Analysis with JMPLast week I had the pleasure to participate in the first <a href="http://www.jmp.com/about/events/summit-europe/live.shtml" target="_blank">European JMP Discovery Summit</a>. As part of this I was able to give a talk on our JMP-Add-In extending JMPs capabilities in spatial data analysis. Now let me show you how to use the Add-In to geocode some addresses in JMP!<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiGNcXD4H5YNqVnleTdTEjhYcon7051vu-_gnZMucx3kjgK1-nipbIp3djbxQNzuN0h7cxcs_OZSCXM9kb1dyso-dU9vDwdSELnQBB9NVsoI7XHCKr1rTB8xdN2QMSubblya89LMfDyfSc/s1600/Header.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiGNcXD4H5YNqVnleTdTEjhYcon7051vu-_gnZMucx3kjgK1-nipbIp3djbxQNzuN0h7cxcs_OZSCXM9kb1dyso-dU9vDwdSELnQBB9NVsoI7XHCKr1rTB8xdN2QMSubblya89LMfDyfSc/s1600/Header.png" height="160" width="640" /></a></div>
<br />
<a name='more'></a>As the conference was located at Brussels and the talk was all about mapping the right example was not far away: Let's do some sightseeing, of course!<br />
<br />
If you don't know what the <a href="http://www.brussels.be/artdet.cfm/5316" target="_blank">Brussels Comic Route</a> is all about you'll get some impressions <a href="http://www.areyoumad.net/bd_en_ville/asterix" target="_blank">here</a>. Basically you'll find some walls of buildings in the city center of brussels covered with large-scaled comics.<br />
<br />
First of all check <a href="http://en.wikipedia.org/wiki/Brussels'_Comic_Book_Route" target="_blank">wikipedia</a> to see the list of all addresses. We can easily import those addresses by using the <i><b>Internet Open ...</b></i> -command in JMP. Go to <i><b>File -> Internet Open</b></i> in JMP and paste the Wikipedia-URL into the dialog.<br />
<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhvTS_fgyhXSpiZpFs-YH9WeHLczlPE60coDcez-oIpooCBo4yUY356tWJFV4oWH992D5MAXgfV3wiY-H4FKR0FT1zijqSA-Gz7BIrtaR7dEfNe2jpMH5sw17_LM9jKyoISvNlzTF3XKoc/s1600/wikipedia.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhvTS_fgyhXSpiZpFs-YH9WeHLczlPE60coDcez-oIpooCBo4yUY356tWJFV4oWH992D5MAXgfV3wiY-H4FKR0FT1zijqSA-Gz7BIrtaR7dEfNe2jpMH5sw17_LM9jKyoISvNlzTF3XKoc/s1600/wikipedia.png" height="318" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Addresses for the Brussels Comic Tour on Wikipedia</td></tr>
</tbody></table>
Select only the first table in the next step. Now you should see the list of all comic-walls in Brussel's city center. It should look like this:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiglB-rfPVh9qNqUmI9bBeXCdy3s_Emw0P6IGvOpbJZiBgnKkchKZHnYr_wQsg7o-5lfGXaovcwkfLBzjx0QcyYc3WHLLPu60ZEpR4V97eTEaDCYz167ymyGXP3Ys7Mmfsa5Zpz0WLomsc/s1600/Brussels+Comic+Tour+Data.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiglB-rfPVh9qNqUmI9bBeXCdy3s_Emw0P6IGvOpbJZiBgnKkchKZHnYr_wQsg7o-5lfGXaovcwkfLBzjx0QcyYc3WHLLPu60ZEpR4V97eTEaDCYz167ymyGXP3Ys7Mmfsa5Zpz0WLomsc/s1600/Brussels+Comic+Tour+Data.PNG" height="390" width="640" /></a></div>
<br />
The next step for geocoding is to slightly chance the address column. Sadly the add-in is not a wizard thus it will not know in which city those streets are. I used a formula to add the word Brussels to each address.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpTFkemSUOYcaE0UhMezh97JhjRIAen8WIe7yhZ8NEHAlkKafE-YFQ8l6GIAQHyqBz7EfpBcvfBF0mn1EXQsTllMONJR8QHEY-EEx-Qv7ghQI79f1y4kyksV4d6oJDgkA7EsBP6LYK_Dk/s1600/formula.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpTFkemSUOYcaE0UhMezh97JhjRIAen8WIe7yhZ8NEHAlkKafE-YFQ8l6GIAQHyqBz7EfpBcvfBF0mn1EXQsTllMONJR8QHEY-EEx-Qv7ghQI79f1y4kyksV4d6oJDgkA7EsBP6LYK_Dk/s1600/formula.PNG" height="400" width="367" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
Now select <i><b>Add-Ins -> Spatial Data Analysis -> Geocode </b></i>from the main menu.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj5DgSKLCIMg-H11Xopf9lsonvMoKDEMtXIRoC9WMnnXPitzoWMRfFqRdhnbS8FV05-gk1R-ARBOWoxeF0Oc989fviAvm5rv4X9qXXObKfjns4wE6BaHxNVv9bNw7Nos786ZJU2T3EDNBE/s1600/Dialog.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj5DgSKLCIMg-H11Xopf9lsonvMoKDEMtXIRoC9WMnnXPitzoWMRfFqRdhnbS8FV05-gk1R-ARBOWoxeF0Oc989fviAvm5rv4X9qXXObKfjns4wE6BaHxNVv9bNw7Nos786ZJU2T3EDNBE/s1600/Dialog.PNG" height="146" width="400" /></a></div>
<br />
The newly created address-column will be used on the right hand side of the dialog. Loading the coordinates from the internet (using R and the Google Maps API) might take a moment. But soon you should see two new columns in the data set: <b>latitude</b> and <b>longitude</b> for each address.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEha7YvKrWkfi5VXVomNznLcI8bYYQV1pN8NVZTZeKHz7Ka5BaLXh_oTJiHNtnPby4SLIedXtU_pB0dfG9yAbfmEPJ_9DWhUr9CeYxwK9e5rl46FvJgAjGHqXcXrczkWEuKsBAaSeAfp_yI/s1600/Brussels+Comic+Tour+Data+Geocoded.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEha7YvKrWkfi5VXVomNznLcI8bYYQV1pN8NVZTZeKHz7Ka5BaLXh_oTJiHNtnPby4SLIedXtU_pB0dfG9yAbfmEPJ_9DWhUr9CeYxwK9e5rl46FvJgAjGHqXcXrczkWEuKsBAaSeAfp_yI/s1600/Brussels+Comic+Tour+Data+Geocoded.PNG" height="394" width="640" /></a></div>
<br />
Now it's easy to present all comic-walls in a map. Just use the graph builder with a background roadmap!<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh6DSCUwtXXBD6gW6yrgY8j99C2lfBQAHBliAvalZY15W2QkIPBs-04FVe57IIrTYtn7hNEIqapWNNdvfkD_qgbSuN4_X7hvAT52nHyfy3pl3FLwNS-72DH3J7aQLRW94Qz13CaKHbn1EQ/s1600/map1.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh6DSCUwtXXBD6gW6yrgY8j99C2lfBQAHBliAvalZY15W2QkIPBs-04FVe57IIrTYtn7hNEIqapWNNdvfkD_qgbSuN4_X7hvAT52nHyfy3pl3FLwNS-72DH3J7aQLRW94Qz13CaKHbn1EQ/s1600/map1.PNG" height="608" width="640" /></a></div>
<br />
If you think this map is a bit messy ... next time I'll show you how to do this in JMP:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiQkqx3tGSw1hHruV6cjrKKkeV0a73i5i6zRXIrnaHKE5gWZ4CUmKLgY4Tuz3C6EOLLSFjXWCYq5DEAM8CQo10j-whCx6dgNu8ZaVEpcZiWP_h795LV2mepJrHUhKQvepIlUYIPtYnQuUs/s1600/map2.PNG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiQkqx3tGSw1hHruV6cjrKKkeV0a73i5i6zRXIrnaHKE5gWZ4CUmKLgY4Tuz3C6EOLLSFjXWCYq5DEAM8CQo10j-whCx6dgNu8ZaVEpcZiWP_h795LV2mepJrHUhKQvepIlUYIPtYnQuUs/s1600/map2.PNG" height="608" width="640" /></a></div>
<br />
If you liked this, don't hesitate to download the Add-In <a href="https://community.jmp.com/docs/DOC-7331" target="_blank">here</a> (be aware that you will need <a href="http://www.r-project.org/" target="_blank">R</a> together with the two packages <a href="http://cran.r-project.org/web/packages/TSP/index.html" target="_blank">TSP</a> and <a href="http://cran.r-project.org/web/packages/ggmap/index.html" target="_blank">ggmap</a>).<br />
<br />
All feedback is welcome at <a href="mailto:sebastian.hoffmeister@statcon.de">sebastian.hoffmeister@statcon.de</a> or <a href="https://twitter.com/STATCON" target="_blank">@statcon</a> on Twitter!<br />
<br />
<br />Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com2tag:blogger.com,1999:blog-276389194507638774.post-25893407333273366312014-10-24T11:04:00.001+02:002014-10-24T11:04:34.089+02:00Start EViews Scripting - NOW!Three days ago I gave my talk on "EViews Scripting and Addin-Development" at the <a href="http://statistiksoftware.blogspot.de/2014/10/review-eviews-user-meeting-2014-in.html" target="_blank">First EViews User Meeting</a>. As I think there is lot of unused potential I'll try to increase the reach of my talk by posting here again.<br />
<br />
<a name='more'></a>My point is, that we as the community of EViews-Users can't expect the EViews developer team to do all the work. There will always be some features that we desire but that are currently not available. Luckely EViews offers such a powerful scripting language and we are often able to implement these features by ourselves. Of course the best part is that it is so easy to share the results of that by creating an addin for EViews based on your code.<br />
<br />
In my talk I explained the whole process of going from a simple idea to the final addin. Starting point was the wish to easily create graphs like that:<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEicyCla1kgi7-U1HvBue7w0M0So2V3unu6LkUGU0f6_pPIxiA3REK7rh2Jr7mJk6KawYIdbrza34dPn02gKk9jj6KSSIl5FGSEWrymd1jEZfviKk5sCz1Y7kquJvoDf9kC0MEkQXfG6DH8/s1600/IMG_1272.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEicyCla1kgi7-U1HvBue7w0M0So2V3unu6LkUGU0f6_pPIxiA3REK7rh2Jr7mJk6KawYIdbrza34dPn02gKk9jj6KSSIl5FGSEWrymd1jEZfviKk5sCz1Y7kquJvoDf9kC0MEkQXfG6DH8/s1600/IMG_1272.JPG" height="300" width="400" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">My kind of Forecast Graph</td></tr>
</tbody></table>
To get an idea how I was able to solve that problem, just download the attached <a href="http://statcon.de/blog/addin.aipz" target="_blank">addin-file</a>. You can open the file by Unzipping it and the whole program code is available in there. Additionally you can find the slides of the talk <a href="http://www.statcon.de/blog/Scripting%20and%20Addins.pdf" target="_blank">here</a>!<br />
<br />
If there are any questions, don't hesitate to ask in here or via Twitter: <a href="https://twitter.com/StatconConsult" target="_blank">@statconConsult</a>!Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-70786099257474825582014-10-23T14:47:00.001+02:002014-10-24T11:02:52.995+02:00Review: EViews User Meeting 2014 in FrankfurtTwo days of presentations, exchange of ideas and networking are over. For all who have missed it: I'm talking about the first <a href="http://www.statcon.de/eviews_usermeeting_2014_88_de.html" target="_blank">EViews User Meeting</a> in Frankfurt. Organized by <a href="http://www.ihs.com/de/de/index.aspx?scs=1" target="_blank">IHS Germany</a> and <a href="http://www.statcon.de/" target="_blank">STATCON</a> <span style="background-color: white; color: #222222; font-family: arial, sans-serif; font-size: 13px;">the User Meeting brought users from several countries together: Austria, Poland and Germany.</span> <span style="color: #222222; font-family: arial, sans-serif; font-size: 13px;">Under the general topic of</span><span style="color: #222222; font-family: arial, sans-serif; font-size: 13px;"> </span><b style="color: #222222; font-family: arial, sans-serif; font-size: 13px;"><span style="color: #3d85c6;">"Global Commodity Markets - Scenarios, Prices and Forecasts"</span></b><span style="color: #222222; font-family: arial, sans-serif; font-size: 13px;"> </span><span style="color: #222222; font-family: arial, sans-serif; font-size: 13px;">many very informative, helpful talks were presented, covering the range from scientific methodology, technical advices & programming and of course: case-studies!</span><br />
<div style="color: #222222; font-family: arial, sans-serif; font-size: 13px;">
The share of thought about how to best utilize EViews within individual organizations as well as the use of different types of econometric models was the most stimulating element across the two days.<br />
<br /></div>
<a name='more'></a><div class="separator" style="clear: both; text-align: center;">
</div>
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhjOf5vANEfXElNEtPnj491J3Hxes-_4Ix2Sll85ivgGimQqAC9b6m3ED3VOkMcIav2zt56RCyq1oMxkMkYPbgDWX7PIlYu4mD5AQlPJN0yQxb53IXiDR2Q4_Yc7MtQv6DyQILR4GYUALA/s1600/B0iKjk9IUAAAvfB.jpg-large.jpeg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhjOf5vANEfXElNEtPnj491J3Hxes-_4Ix2Sll85ivgGimQqAC9b6m3ED3VOkMcIav2zt56RCyq1oMxkMkYPbgDWX7PIlYu4mD5AQlPJN0yQxb53IXiDR2Q4_Yc7MtQv6DyQILR4GYUALA/s1600/B0iKjk9IUAAAvfB.jpg-large.jpeg" height="400" width="298" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Bertram Schäfer at his talk</td></tr>
</tbody></table>
<b>Day 1</b> started with a talk from Prof. Achim Wübker sharing his ideas about "Forecasting Electricity Rates via EViews incorporating Political Decisions". Other highlights were Dr. Margolis presentation on risk-modelling for industries and companies and Christian Borgmann giving an insight into the "Importance of Energyprice- and gasprice-forecasts for Local Energy Suppliers". The last presentation of the day was mine about <a href="http://www.statcon.de/blog/Scripting%20and%20Addins.pdf" target="_blank">Scripting and Addin-Developement in EViews</a>.<br />
<br />
While the official part was over at that point, we had the chance to continue with the more informal part at the conference dinner. I can't recommend the "Ariston"-restaurant in Frankfurt enough. They provided great food in a fantastic atmosphere.<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiFYi6NDpPZ2DH8vHRADmXgEuLV3ZhdZuSi7Lt1NhDpLdmFpGxiZvZlaO_UHKSTC3ZA-JTW_O7t-GnIGnoydsUh2W7IHrmMQrAz3FODd3hSAHEHlHDZTilmleLWBD3c5ZykDSt3bNu8PYQ/s1600/B0iN4DbCcAEwNQJ.jpg-large.jpeg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiFYi6NDpPZ2DH8vHRADmXgEuLV3ZhdZuSi7Lt1NhDpLdmFpGxiZvZlaO_UHKSTC3ZA-JTW_O7t-GnIGnoydsUh2W7IHrmMQrAz3FODd3hSAHEHlHDZTilmleLWBD3c5ZykDSt3bNu8PYQ/s1600/B0iN4DbCcAEwNQJ.jpg-large.jpeg" height="240" width="320" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Me: Talking about Scripting in EViews</td></tr>
</tbody></table>
We can't talk about <b>day 2</b> of the conference without mentioning Bertram Schäfers talk about "Modelling Energy-Prices using Weather Data". In the end our chair Prof. Wübcker had to stop the discussions to make sure that the other speaker will have their fair share of time, too. One of those speakers was Dr. Goers giving us all an idea what the "MOVE"-project is all about and what kind of models he uses to make predictions for the austrian economy. Last but not least Boris Fuks from IHS took his time to introduce the GLM (Global Link Model) which seems to be a great tool to get advantage from the power of EViews without needing to care for how to use EViews itself.<br />
<br />
All in all we had great two days and I want to thank all the speakers, organizers (especially Mr. Tröscher from IHS and Claudia Walber from STATCON) and of course all of the participants for making this kind of event possible! I hope to see all of you - and even more - next year for the Second EViews User Meeting!<br />
<br />
If you want to make sure to be informed:<br />
- Follow us on twitter: <a href="https://twitter.com/STATCON" target="_blank">@statcon</a> and <a href="https://twitter.com/StatconConsult" target="_blank">@statconConsult</a><br />
- Ask via mail: <a href="mailto:vertrieb@statcon.de">vertrieb@statcon.de</a> and <a href="mailto:consult@statcon.de">consult@statcon.de</a><br />
<br />
Take care!<br />
SebastianSebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-15415504540725806192014-07-29T15:51:00.002+02:002014-07-29T15:52:10.907+02:00Noticable Pieces (1): Beer Blends - A Taste for Mixtures<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgSU1tlW1wzEihtQagXKhRHeHFwhLXGgVmC6gH6uLXzDRWJsQXc2Hrq1Do7s4LSzviJSrdCbNIjK-MYHrk1KKGCag4f4NHYkT_MkS7hS28DVTHHOdlsczkNvNSpgAy9UqmvxktzYhmdCr8/s1600/fattymattybrewing_Fatty_Matty_Brewing_-_Beer_Mug_Icon.png" imageanchor="1" style="clear: right; float: right; margin-bottom: 1em; margin-left: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgSU1tlW1wzEihtQagXKhRHeHFwhLXGgVmC6gH6uLXzDRWJsQXc2Hrq1Do7s4LSzviJSrdCbNIjK-MYHrk1KKGCag4f4NHYkT_MkS7hS28DVTHHOdlsczkNvNSpgAy9UqmvxktzYhmdCr8/s1600/fattymattybrewing_Fatty_Matty_Brewing_-_Beer_Mug_Icon.png" height="200" width="200" /></a></div>
<br />
In my previous article I mentioned the beer-tasting experiment as one of the highlights of the <a href="http://statistiksoftware.blogspot.de/2014/07/like-cupcakes-go-to-5th-european-design.html" target="_blank">5th European DoE User Meeting</a>.<br />
<br />
Sadly I was not around to see the analysis of the data - but luckily Andrew and Paul of PrismTC wrote an excellent article on their website.<br />
<br />
So everyone who is either interested in beer and/or mixture experiments: Read <a href="http://www.prismtc.co.uk/beer-blends/" target="_blank">this</a>!Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-68621268187140355922014-07-17T09:01:00.000+02:002014-07-17T11:10:55.018+02:00Like Cupcakes? Go to the 5th European Design of Experiments User Meeting!<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg3aqxeJaz4FbNTkwyPDgZ30_mMdsdcW1r77YrFeIZiTx4qNSp8K_Ru5K_F2PwSjW8UQXLxOr94irBTI71isZocEStEWOrjW68b117ambk79HIUQUYnd52cwpRqME_XHzFrLylkzUKBz9M/s1600/Cupcakes.jpg" imageanchor="1" style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><br />
</a></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg5w6wF57EhIDEV5QnD3pqBR3AR94m8GlINJwHelEEA27eEoJq01hAFDSeg4yoFglumU1-EoURK0nbU2kXgIb7kTDj1u5W1wPCLuTJSasruzMY3C3nCLGJ_pQnk97W7kgw9Bxp7Ce_TTXs/s1600/photo.JPG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg5w6wF57EhIDEV5QnD3pqBR3AR94m8GlINJwHelEEA27eEoJq01hAFDSeg4yoFglumU1-EoURK0nbU2kXgIb7kTDj1u5W1wPCLuTJSasruzMY3C3nCLGJ_pQnk97W7kgw9Bxp7Ce_TTXs/s1600/photo.JPG" height="240" width="320" /></a></div>
<br />
Today I want to share a few of my impressions from the <a href="http://www.prismtc.co.uk/post-event-information/" target="_blank">5th Design of Experiments User Meeting</a> in Cambridge. After being organized by <a href="http://www.statcon.de/" target="_blank">Statcon</a> for the last three times (Berlin, Luzern -in cooperation with<a href="http://cq.be/en/" target="_blank"> CQ-Consultancy</a> -, and <a href="http://www.statcon.de/4_europaeisches_doe_user_meeting_in_wien_70_de.html" target="_blank">Vienna</a>), <a href="http://www.prismtc.co.uk/" target="_blank">PRISMTC</a> from the UK was the host in Cambridge.<br />
<br />
<a name='more'></a><br />
<br />
Before I start: What is the "DoE User Meeting" all about? Well it is an incredible event to get in touch with people using DoE from all over Europe. Actually "from all over the world" might be more precise. This includes not only statisticians but for the largest part practitioners from all areas of the industry.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjeQuNrVwL_o0jHTCEtP-HmJAz91FvTbLOAWsKCOJH4pcWOI-WtQFBjGJ9CIMGPcdTjYy166pLNEqK6sNmqLWoPM-Zp66ZJaXqWU6AiA_QzNBC7bkm0q-xksiQc6oZOzMNwBGCINh9ueYA/s1600/Downing+College.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjeQuNrVwL_o0jHTCEtP-HmJAz91FvTbLOAWsKCOJH4pcWOI-WtQFBjGJ9CIMGPcdTjYy166pLNEqK6sNmqLWoPM-Zp66ZJaXqWU6AiA_QzNBC7bkm0q-xksiQc6oZOzMNwBGCINh9ueYA/s1600/Downing+College.jpg" height="424" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Howards Building in Downing Collage - Our Accommodation </td></tr>
</tbody></table>
The Statcon-team - being Bertram Schäfer (CEO), Claudia Walber (Sales Manager) and me - arrived on monday. Thus I was able to take my camera and catch some impressions before the reseller meeting started on tuesday morning.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjuudIuWNCLOUk-MMys_Gsvt4u7OSHxCdkE6PHI2HJA0RGAJwHO9ghVELu2aq_1K8krHRnZ8ZC-MYx1KBU2UtPDtwOACYdOOPfIWo_ULDsRIIeGmK3bWUs2gyv5t3SynN2gDMQDBffg3NI/s1600/DSC06211.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjuudIuWNCLOUk-MMys_Gsvt4u7OSHxCdkE6PHI2HJA0RGAJwHO9ghVELu2aq_1K8krHRnZ8ZC-MYx1KBU2UtPDtwOACYdOOPfIWo_ULDsRIIeGmK3bWUs2gyv5t3SynN2gDMQDBffg3NI/s1600/DSC06211.jpg" height="424" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Cambridge at Night</td></tr>
</tbody></table>
This was a closed event for the sales departments of PRISMTC (UK), Ritme (France), CQ-Consultancy (Belgium), Statcon (Germany) and of course StatEase. The whole day was full of discussions, ideas and networking.<br />
<br />
Wednesday was the first official day of the User Meeting. In two parallel workshops beginners could get a great introduction to DoE from Paul Nelson and Andrew Macpherson while the advanced DoE-users could learn about "Robust Designs and Tolerance Analysis" from Mark Anderson and Pat Whitcomb.<br />
<br />
Thursday started with the talk of Mark Anderson (Cofounder of StatEase) giving an overview over important new features of Design Expert in version 9. It was a must-hear for everyone using version 9, as it was not only about the major features (Split-Plots) but showed a huge number of small improvements. The time passed quickly with a long number of fascinating talks about applications of DoE in the real world.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgCfGO111ij7CjjHFYgPGL75vc-uwlAnMTpTPCeeKQflC2Cx5IAb8ipaxN5a6tpOEMbN5ryRyeFVYnKpQuzuP7RA5njQCEJ39FxVEWpnnCDt0Ygu7lIsCyhRy9eVmw2IGqpwVOp0YY9tp8/s1600/Eagles+Pub.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgCfGO111ij7CjjHFYgPGL75vc-uwlAnMTpTPCeeKQflC2Cx5IAb8ipaxN5a6tpOEMbN5ryRyeFVYnKpQuzuP7RA5njQCEJ39FxVEWpnnCDt0Ygu7lIsCyhRy9eVmw2IGqpwVOp0YY9tp8/s1600/Eagles+Pub.jpg" height="424" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">The Eagle Pub</td></tr>
</tbody></table>
Of course the highlight was still coming. Right after the talks we split up into four groups to experience the scientific history-tour of cambridge. <i>My</i> tour guide Toni did an fantastic job allowing us to feel the touch of history while walking through the historic center of cambridge.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhWZz9Xw6tBL3Sd_E1pUutxVCKU2-A7yg0Xpx0wdU4tlMFhri4uQW9dpcx1D_acP8SXCuoXr-wn4t8IhvINvM-hd1pW2sLxMGOLu6VmtK_6EoOs0cVIPZvUgfc_SmDRV4z5X_TYs7iPxBc/s1600/dinner.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhWZz9Xw6tBL3Sd_E1pUutxVCKU2-A7yg0Xpx0wdU4tlMFhri4uQW9dpcx1D_acP8SXCuoXr-wn4t8IhvINvM-hd1pW2sLxMGOLu6VmtK_6EoOs0cVIPZvUgfc_SmDRV4z5X_TYs7iPxBc/s1600/dinner.JPG" height="316" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Conference Dinner</td></tr>
</tbody></table>
The tour ended in Magdalene College where Paul and Andrew started a beer tasting experiment. I'm regretting it so much that I could not stay long enough to hear the results of this study on friday afternoon. The day ended with a delicious candle-light-dinner.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhxidyoVim1Vz3SosEP9o0kUE8bwvBlEi8YvDqcn2wYE2azYBkB6lujSQigNfWDYJ_foTwW95kjqFdGZExt35Ct3JXEBfExAuDa821D4UNZweKEu2CTLTBhRuuACtFhCmm9wpNBgJfw9Ww/s1600/Venue.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhxidyoVim1Vz3SosEP9o0kUE8bwvBlEi8YvDqcn2wYE2azYBkB6lujSQigNfWDYJ_foTwW95kjqFdGZExt35Ct3JXEBfExAuDa821D4UNZweKEu2CTLTBhRuuACtFhCmm9wpNBgJfw9Ww/s1600/Venue.jpg" height="360" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Venue</td></tr>
</tbody></table>
Friday started with Pat Whitcomb (Cofounder of StatEase) discussing the pros and cons of split-plot designs. Right after him I gave my best to show why we need split-plots designs so badly and how we apply them in DesignExpert 9. As a showcase I used the example of baking cupcakes. Of course: After talking so much about cupcakes I had to provide some to the audience!<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEihFAK9iXmA650RoCGxATmX6mz_1xSkzjipyquakoGvZXwBGaBPflgOhQuqmMDs5w1ZQYHyIAUNl9AQLJqU7cjk3KCCnLjiKdJwfonUtMYNFwxPt8mBnOiCHmRqdwOzwUzCiMCGbw2XM3Q/s1600/Cupcakes.jpg" height="320" width="246" /><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjnjR_v8vyy-tyQkK2fXfvZc80nnWC8FKOxjXfVqDtWeiw9GCUnM64fiEHpEMk4rPMOYtXc4BbShubJ5AILzdLWN7JgCygF2q8TD7izjrBWLNcZMy4bImAVoNLoAa8IMlMYVHCur3JHWIQ/s1600/P2+-+Data+Collection+Protocol.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjnjR_v8vyy-tyQkK2fXfvZc80nnWC8FKOxjXfVqDtWeiw9GCUnM64fiEHpEMk4rPMOYtXc4BbShubJ5AILzdLWN7JgCygF2q8TD7izjrBWLNcZMy4bImAVoNLoAa8IMlMYVHCur3JHWIQ/s1600/P2+-+Data+Collection+Protocol.png" height="300" width="400" /></a><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhzMHGrG1O7_NBwPHU2VEqYjl6K2XXkEWvuuQh91CP7uA5Oj-GRZX8zsREoAxYO64mnNgWw-QykHMueYU2a0MLzZt8ybAcfFgenZXImg7Qg5p67rU8N3HY1wkCTMakC5Z3UEHi2ohIiZqM/s1600/Speakers.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"></a></td></tr>
</tbody></table>
<br />
Again multiple great talks about applied DoE followed! Finally I would like to conclude with saying thank you!<br />
<br />
Thank you <b>PRISMTC</b> for doing an incredible good job in organizing a perfect event!<br />
<br />
Thank you <b>StatEase</b> for caring so much for your customers and giving us the change to interact with you!<br />
<br />
Thank you all <b>participants</b> and <b>speakers</b> for enabling an event like this with that many options for networking and sharing of ideas!Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-16696954479513542712014-05-05T17:00:00.000+02:002014-08-21T09:08:13.486+02:00Control 2014<div style="text-align: center;">
Die letzten Arbeiten an unserem Messestand 221, Halle 3 werden durch unseren Chef Herrn Bertram Schäfer verrichtet.</div>
<div style="text-align: center;">
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiD1D2ZnUQZvAGOdjTK295ez6vE-0a9YpMYEWOXCt0BRTEjT5e40rye7HFzowYRL5UqdmS7ZIY5K8SpBv3AykZBQtHFcpS7obAiJI0ACW3WvNc9mPIZDr_DgXffapC95rCFOZn8u3Ua_Hc/s1600/photo01gr.JPG" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiD1D2ZnUQZvAGOdjTK295ez6vE-0a9YpMYEWOXCt0BRTEjT5e40rye7HFzowYRL5UqdmS7ZIY5K8SpBv3AykZBQtHFcpS7obAiJI0ACW3WvNc9mPIZDr_DgXffapC95rCFOZn8u3Ua_Hc/s1600/photo01gr.JPG" height="240" width="320" /></a></div>
<div style="text-align: center;">
<br /></div>
<div style="text-align: center;">
Besuchen Sie uns, um das Ergebnis zu begutachten.</div>
Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com070629 Stuttgart, Deutschland48.689385599999987 9.210137499999973448.647455599999986 9.1294564999999732 48.731315599999988 9.2908184999999737tag:blogger.com,1999:blog-276389194507638774.post-68085318385800296862014-02-07T14:42:00.000+01:002014-02-07T14:57:40.970+01:00Der Geschmack des Mittelwerts<div style="text-align: justify;">
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhDJmX5B69fVHk11yc9ESzjBtdM4RChXeRDc7MW89hcudF15W00nl3pNKC_rUrGC70hQrAyGZwBKKfnzthb3n-atp1_7-rwxXWgjZuTsTrK59pp-gBJdyusunxdRViUdFhvcWgZEFrya_M/s1600/cupcakes.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhDJmX5B69fVHk11yc9ESzjBtdM4RChXeRDc7MW89hcudF15W00nl3pNKC_rUrGC70hQrAyGZwBKKfnzthb3n-atp1_7-rwxXWgjZuTsTrK59pp-gBJdyusunxdRViUdFhvcWgZEFrya_M/s1600/cupcakes.png" height="273" width="320" /></a></div>
<span style="font-family: Verdana, sans-serif;">Wer gerne Max und Caroline in <a href="http://www.cbs.com/shows/2_broke_girls">2 Broke Girls</a> bei ihrem endlosen Bestreben ein Cupcake-Imperium aufzubauen verfolgt, der kommt nicht umhin selber mal den Versuch zu wagen. <i>Leider </i>ist die Menge an verfügbaren Rezepten recht umfangreich (<a href="http://www.chefkoch.de/rs/s0/cupcake/Rezepte.html">Chefkoch.de liefert 385 Rezepte</a>), so dass es wohl zumindest den einen oder anderen Entscheidungsbaum(-kuchen) bräuchte um das beste Rezept zu finden. Wer sich zwischen <a href="http://www.chefkoch.de/rezepte/1874091304583166/Raffaelo-Cupcakes.html">Rafaello-Cupcake</a>, <a href="http://www.chefkoch.de/rezepte/1729791282109395/Schwarzwaelder-Kirsch-Cupcakes.html">Schwarzwälder-Kirsch</a> oder gar dem <a href="http://www.chefkoch.de/rezepte/1993221322834906/Schokoladenmuffins-mit-einem-Marshmellowhut.html">Schockoladenmuffin mit Marshmellowhut</a> wählen muss hat auf jeden Fall gravierende Erste-Welt-Probleme.</span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"></span><br />
<a name='more'></a><span style="font-family: Verdana, sans-serif;"><br /></span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;">Deshalb stecke ich mir erst mal kleinere Ziele: Ich will ein gutes Basis-Rezept finden, auf dessen Grundlage ich meine eigenen geschmacklichen $\sigma$^2en entwickeln kann. In guter Statistikermanier gehe ich also folgendermaßen vor: </span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><br /></span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><b>Ad Primum </b>Sammel die besten Cupcake-Rezepte.</span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><br /></span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><b>Ad Secundum</b> Erforsche was diese Rezepte gemeinsam haben.</span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><br /></span></div>
<div style="text-align: justify;">
<span style="font-family: Verdana, sans-serif;"><b>Postremo</b> Erstelle ein Rezept auf Basis der bis dahin gemachten Erfahrungen und backe es nach!</span></div>
<div style="text-align: justify;">
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
<span style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh4CBN9MShgFpojYFZCuOrUvC_5hFaPCi-k4vwVaU55VyN-4cEWnVpEGfAdPwODv8neK4dbFOJR60KpW_QnOhNyOuKiL5A3EJajRO561TgZhlVx62OxNwd8X0nPzOcEtywn4eTZSlR6oYg/s1600/A_-_Relacio_1646.jpg" /><span style="font-size: large;"><b>d Primum</b></span></span></div>
<h3>
<span style="color: #674ea7;"><br /></span></h3>
<h3>
</h3>
<div>
<span style="color: #674ea7;"><br /></span>
<br />
<br />
<br />
<br />
<br /></div>
<div>
<div style="text-align: center;">
<table border="1" cellspacing="0" style="text-align: center;">
<caption align="bottom" class="captiondataframe"></caption>
<tbody>
<tr><td><table border="0" class="dataframe">
<tbody>
<tr class="firstline">
<th> </th>
<th>Zucker </th>
<th>Butter </th>
<th>Eier </th>
<th>Backpulver </th>
<th>Salz </th>
<th>Mehl </th>
<th>Milch </th>
<th>Temperatur </th>
<th>Zeit</th>
</tr>
<tr>
<td class="firstcolumn">Schwarzwaelder-Kirsch
</td>
<td class="cellinside">150
</td>
<td class="cellinside">125
</td>
<td class="cellinside">2.0
</td>
<td class="cellinside">1.00
</td>
<td class="cellinside">1
</td>
<td class="cellinside">175
</td>
<td class="cellinside">0.0
</td>
<td class="cellinside">180
</td>
<td class="cellinside">20
</td></tr>
<tr>
<td class="firstcolumn">Schokomuffin
</td>
<td class="cellinside">200
</td>
<td class="cellinside">125
</td>
<td class="cellinside">2.3
</td>
<td class="cellinside">1.80
</td>
<td class="cellinside">0
</td>
<td class="cellinside">194
</td>
<td class="cellinside">194.0
</td>
<td class="cellinside">190
</td>
<td class="cellinside">20
</td></tr>
<tr>
<td class="firstcolumn">Nutella Cupcake
</td>
<td class="cellinside">280
</td>
<td class="cellinside">80
</td>
<td class="cellinside">2.0
</td>
<td class="cellinside">3.00
</td>
<td class="cellinside">2
</td>
<td class="cellinside">200
</td>
<td class="cellinside">240.0
</td>
<td class="cellinside">180
</td>
<td class="cellinside">15-20
</td></tr>
<tr>
<td class="firstcolumn">Kruemelmonster
</td>
<td class="cellinside">200
</td>
<td class="cellinside">125
</td>
<td class="cellinside">3.0
</td>
<td class="cellinside">3.50
</td>
<td class="cellinside">0
</td>
<td class="cellinside">300
</td>
<td class="cellinside">125.0
</td>
<td class="cellinside">180
</td>
<td class="cellinside">20-25
</td></tr>
<tr>
<td class="firstcolumn">Schoko-Kuesst-Himbeer
</td>
<td class="cellinside">150
</td>
<td class="cellinside">120
</td>
<td class="cellinside">2.0
</td>
<td class="cellinside">1.50
</td>
<td class="cellinside">0
</td>
<td class="cellinside">100
</td>
<td class="cellinside">150.0
</td>
<td class="cellinside">180
</td>
<td class="cellinside">20
</td></tr>
<tr>
<td class="firstcolumn">Cupcale Schafe mit Marshmellow Frosting
</td>
<td class="cellinside">120
</td>
<td class="cellinside">120
</td>
<td class="cellinside">3.0
</td>
<td class="cellinside">3.00
</td>
<td class="cellinside">0
</td>
<td class="cellinside">200
</td>
<td class="cellinside">0.0
</td>
<td class="cellinside">180-200
</td>
<td class="cellinside">20-25
</td></tr>
<tr>
<td class="firstcolumn">Schoko Cupcakes
</td>
<td class="cellinside">188
</td>
<td class="cellinside">125
</td>
<td class="cellinside">2.5
</td>
<td class="cellinside">0.99
</td>
<td class="cellinside">1
</td>
<td class="cellinside">125
</td>
<td class="cellinside">163.0
</td>
<td class="cellinside">175
</td>
<td class="cellinside">25
</td></tr>
<tr>
<td class="firstcolumn">Cupcakes
</td>
<td class="cellinside">125
</td>
<td class="cellinside">125
</td>
<td class="cellinside">2.0
</td>
<td class="cellinside">2.00
</td>
<td class="cellinside">1
</td>
<td class="cellinside">200
</td>
<td class="cellinside">0.2
</td>
<td class="cellinside">180
</td>
<td class="cellinside">20-25
</td></tr>
<tr>
<td class="firstcolumn">Raffaelo Cupcakes
</td>
<td class="cellinside">150
</td>
<td class="cellinside">75
</td>
<td class="cellinside">1.5
</td>
<td class="cellinside">1.95
</td>
<td class="cellinside">1
</td>
<td class="cellinside">188
</td>
<td class="cellinside">150.0
</td>
<td class="cellinside">175
</td>
<td class="cellinside">20
</td></tr>
<tr>
<td class="firstcolumn">Zitronen Cupcake
</td>
<td class="cellinside">90
</td>
<td class="cellinside">120
</td>
<td class="cellinside">2.0
</td>
<td class="cellinside">2.00
</td>
<td class="cellinside">1
</td>
<td class="cellinside">190
</td>
<td class="cellinside">110.0
</td>
<td class="cellinside">175
</td>
<td class="cellinside">20
</td></tr>
</tbody>
</table>
</td></tr>
</tbody></table>
</div>
</div>
<div>
<br />
<br />
<div>
<span style="font-family: Verdana, sans-serif;">Die besten zehn Gebote ... äh ... Rezepte von Chefkoch.de sind schnell gesammelt! Natürlich sind diese Rezepte schon ein wenig bereinigt: Spezielle Zutaten für den besondere Geschmack (Erdbeere, Nutella, ...) wurden entfernt.</span></div>
<span style="font-family: Verdana, sans-serif;"><br /></span>
<span style="font-family: Verdana, sans-serif;">Im nächsten Schritt werden noch die Einheiten (<i>gr, Stück, TL, Prisen, ml</i>) auf eine gemeinsame Einheit <b>Gramm</b> umgerechnet. Für den nächsten Schritt kann man also auf folgende Daten zugreifen.</span><br />
<br />
<div align="center">
<table border="1" cellspacing="0">
<caption align="bottom" class="captiondataframe"></caption>
<tbody>
<tr><td><table border="0" class="dataframe">
<tbody>
<tr class="firstline">
<th> </th>
<th>Zucker </th>
<th>Butter </th>
<th>Eier </th>
<th>Backpulver </th>
<th>Salz </th>
<th>Mehl </th>
<th>Milch</th>
</tr>
<tr>
<td class="firstcolumn">Schwarzwaelder-Kirsch
</td>
<td class="cellinside">0.27
</td>
<td class="cellinside">0.23
</td>
<td class="cellinside">0.18
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.32
</td>
<td class="cellinside">0.00
</td></tr>
<tr>
<td class="firstcolumn">Schokomuffin
</td>
<td class="cellinside">0.24
</td>
<td class="cellinside">0.15
</td>
<td class="cellinside">0.14
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.23
</td>
<td class="cellinside">0.23
</td></tr>
<tr>
<td class="firstcolumn">Nutella Cupcake
</td>
<td class="cellinside">0.31
</td>
<td class="cellinside">0.09
</td>
<td class="cellinside">0.11
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.22
</td>
<td class="cellinside">0.26
</td></tr>
<tr>
<td class="firstcolumn">Kruemelmonster
</td>
<td class="cellinside">0.22
</td>
<td class="cellinside">0.14
</td>
<td class="cellinside">0.16
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.33
</td>
<td class="cellinside">0.14
</td></tr>
<tr>
<td class="firstcolumn">Schoko-Kuesst-Himbeer
</td>
<td class="cellinside">0.24
</td>
<td class="cellinside">0.19
</td>
<td class="cellinside">0.16
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.16
</td>
<td class="cellinside">0.24
</td></tr>
<tr>
<td class="firstcolumn">Cupcake Schafe mit
Marshmellow Frosting
</td>
<td class="cellinside">0.20
</td>
<td class="cellinside">0.20
</td>
<td class="cellinside">0.25
</td>
<td class="cellinside">0.02
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.33
</td>
<td class="cellinside">0.00
</td></tr>
<tr>
<td class="firstcolumn">Schoko Cupcakes
</td>
<td class="cellinside">0.26
</td>
<td class="cellinside">0.17
</td>
<td class="cellinside">0.17
</td>
<td class="cellinside">0.00
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.17
</td>
<td class="cellinside">0.22
</td></tr>
<tr>
<td class="firstcolumn">Cupcakes
</td>
<td class="cellinside">0.22
</td>
<td class="cellinside">0.22
</td>
<td class="cellinside">0.18
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.36
</td>
<td class="cellinside">0.00
</td></tr>
<tr>
<td class="firstcolumn">Raffaelo Cupcakes
</td>
<td class="cellinside">0.23
</td>
<td class="cellinside">0.12
</td>
<td class="cellinside">0.12
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.29
</td>
<td class="cellinside">0.23
</td></tr>
<tr>
<td class="firstcolumn">Zitronen Cupcake
</td>
<td class="cellinside">0.15
</td>
<td class="cellinside">0.19
</td>
<td class="cellinside">0.16
</td>
<td class="cellinside">0.01
</td>
<td class="cellinside">0
</td>
<td class="cellinside">0.31
</td>
<td class="cellinside">0.18
</td></tr>
</tbody>
</table>
</td></tr>
</tbody></table>
<span style="font-family: Verdana, sans-serif;">Jede Zeile enthält die prozentualen Anteile der Zutaten im jeweiligen Rezept.</span><br />
<div style="text-align: left;">
<br /></div>
<div style="text-align: left;">
<img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjpDrVwn1pRqi7xIlzdKdoKLUPt_TYn8OaTwVk4Ws1eddMF_jIeVWJE5nv0DiQgwPf5CImzVR86XBs1FLYQH1WKpfx6sNqCpMPwnfMVSe8FP531SL6oApNc6s_U_0jVVGCABvVa9SlzKv0/s1600/Letter_S_with_Flowers.jpg" /> <span style="color: black; font-size: large;"><b>ecundum: Grafische Datenauswertung</b></span></div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiUiGDZMYKdAHjBBeuslbK8zLE4alc0XeLrRmY-xZjFiQhVKMFLGSTGkVl0gTEf5ePPPlckR-zlwFlQdyzyjSpFTMHUlst7Kr3rgpAeJgKEqyD9FtiDonj41thAsxhpJdPF-fHA6qLzfNQ/s1600/bp_kuchen.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiUiGDZMYKdAHjBBeuslbK8zLE4alc0XeLrRmY-xZjFiQhVKMFLGSTGkVl0gTEf5ePPPlckR-zlwFlQdyzyjSpFTMHUlst7Kr3rgpAeJgKEqyD9FtiDonj41thAsxhpJdPF-fHA6qLzfNQ/s1600/bp_kuchen.png" height="320" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-family: Verdana, sans-serif; font-size: small;">Boxplots der prozentualen Anteile der Zutaten</span></td></tr>
</tbody></table>
<div style="text-align: left;">
<span style="font-family: Verdana, sans-serif;">Salz und Backpulver scheinen mir ein schönes Beispiel für die Verwendung des Variationskoeffizienten als alternatives Maß für die Streuung, gegenüber der Varianz. Aber wie kommt man jetzt auf ein Rezept? Ich nehme einfach den mittleren prozentualen Anteil jeder Zutat und normiere das Ergebnis auf 100%.</span></div>
<br />
<div style="text-align: right;">
<span style="text-align: left;"><br /></span></div>
<div style="text-align: right;">
<span style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiEsl5A_PIzNRPk1_BXfk7TEJ84ldcDUSlwtO9cefHejfMvpvEeKKiym0jGHs407dTdUlUlENAt_qiFD0Nr2ts8gDojs1Q7oLL26q3zein4htWijRcPnD5OMDCz_PWg5cmIQqjXCZyX6wU/s1600/Fancy_Letter_P_Image.jpg" /><b><span style="font-size: large;">ostremo: Das endgültige Rezept</span></b></span><span style="text-align: left;"><br /></span></div>
<div style="text-align: right;">
<span style="text-align: left;"><br /></span></div>
<div style="text-align: right;">
<span style="text-align: left;"><br /></span></div>
<div style="text-align: right;">
<span style="text-align: left;"><br /></span></div>
<div style="text-align: right;">
<span style="text-align: left;"><br /></span></div>
<div style="text-align: left;">
<br /></div>
<div style="text-align: left;">
<span style="font-size: large;"><br /></span></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjvPHTAwFThOZrbDwqxRPfT81MBJ_48YZL9zKV-lCZoQ91Vbq-Bvjmp4XvqqTAK69wqJzpI4paEENV1VawHuTq74nLg4jh6CjMScqefN-SAQ64VlQsN7yTK1nNXdowK8XD9iyd8qTOB1fo/s1600/screenshot_10.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjvPHTAwFThOZrbDwqxRPfT81MBJ_48YZL9zKV-lCZoQ91Vbq-Bvjmp4XvqqTAK69wqJzpI4paEENV1VawHuTq74nLg4jh6CjMScqefN-SAQ64VlQsN7yTK1nNXdowK8XD9iyd8qTOB1fo/s1600/screenshot_10.jpg" height="44" width="640" /></a></div>
<br />
<div style="text-align: left;">
<span style="font-family: Verdana, sans-serif;">Hochgerechnet auf 200 Gramm Teig braucht man:</span></div>
<ul>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">47 Gramm Zucker</span></i></li>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">34 Gramm Butter</span></i></li>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">33 Gramm Ei (~1 kleines Ei)</span></i></li>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">2 Gramm Backpulver (~ 0.5 TL)</span></i></li>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">54 Gramm Mehl</span></i></li>
<li style="text-align: left;"><i><span style="font-family: Verdana, sans-serif;">30 ml Milch</span></i></li>
</ul>
<div>
<div style="text-align: left;">
<span style="font-family: Verdana, sans-serif;">Und das Resultat:</span></div>
</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgyomBhpIvrqClDRpcXM-YHRfDvU2gZ9_1MEUJtE0bphiO_7GeRTUowWwSDNivGZZzThlqV3De9VKPfPO5zd4szSQyvIxVPhmCk3406wzNp31bknnN6hLXBLUCwvFOZesHSk5aGssUxPz8/s1600/kuchen.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgyomBhpIvrqClDRpcXM-YHRfDvU2gZ9_1MEUJtE0bphiO_7GeRTUowWwSDNivGZZzThlqV3De9VKPfPO5zd4szSQyvIxVPhmCk3406wzNp31bknnN6hLXBLUCwvFOZesHSk5aGssUxPz8/s1600/kuchen.png" height="258" width="320" /></a></div>
<div>
<span style="font-family: Verdana, sans-serif;"><br /></span></div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<div>
<div style="text-align: left;">
<span style="font-family: Verdana, sans-serif;">Viel Spass beim nachbacken!</span></div>
</div>
</div>
<div class="separator" style="clear: both; text-align: center;">
</div>
</div>
</div>
Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com2tag:blogger.com,1999:blog-276389194507638774.post-53462474595119486042013-06-14T13:50:00.002+02:002013-06-14T13:52:06.076+02:00Balkendiagramme für Mittelwerte<div style="text-align: justify;">
Im Rahmen der <a href="http://blogs.sas.com/content/sasdach/2013/04/09/leitmotiv-datenvisualisierung-ihre-meinung-zahlt/">Blogparade</a> von SAS zum Thema Datenvisualisierung gibt Betram Schäfer (CEO der Firma Statcon) Hinweise zur Verwendung von Balkendiagrammen.</div>
<div>
<a name='more'></a><div style="text-align: justify;">
<span style="font-family: 'Times New Roman';">Sehr oft werden in einem Balkendiagramm Mittelwerte von Daten einer Gruppe auf der Ordinate (Y-Achse) abgebildet, dabei sollten einige Besonderheiten beachtet werden. </span><span style="font-family: 'Times New Roman';">Beispielhaft sollen die unten dargestellten, erfundenen Daten eines Backexperimentes verwendet werden:</span></div>
<br />
<div style="text-align: justify;">
<br /></div>
</div>
<div style="min-height: 15px;">
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjN16rZjY2IsbarM_3eE9hsiVVhT8a_GksRfcjRRAT12fC3kKForpL60w16d0nX9izHzYXWOvAZ-4kZBBYBuD9Ikkf6mqF2_vaIqEYq2DzyB5Gpu0iabP_N4IOSQTTJYyCzfIE8lQWcMc0/s1600/Daten.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjN16rZjY2IsbarM_3eE9hsiVVhT8a_GksRfcjRRAT12fC3kKForpL60w16d0nX9izHzYXWOvAZ-4kZBBYBuD9Ikkf6mqF2_vaIqEYq2DzyB5Gpu0iabP_N4IOSQTTJYyCzfIE8lQWcMc0/s1600/Daten.jpg" /></a></div>
<div style="text-align: justify;">
<br /></div>
<h3 style="text-align: justify;">
<span style="font-size: small;">Beschreibung des Experimentes:</span></h3>
<div style="text-align: justify;">
Es handelt sich um Daten eines faktoriellen Versuchsplans, bei welchem zwei unterschiedliche Mehlqualitäten (<b style="color: blue;">billig</b> - <span style="color: blue;"><b>teuer</b></span>) bei zwei Backzeiten (<span style="color: blue;"><b>kurz</b></span> - <span style="color: blue;"><b>lang</b></span>) verarbeitet wurden. Alle anderen Komponenten des Rezeptes (Mehlanteil, Wasseranteil, Backtemperatur, ...) blieben unverändert. </div>
</div>
<div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Gemessen wurde der Geschmack anhand einer Gruppe von Testpersonen. Jede Testperson sollte das Brot auf einer Skala von 0 (<span style="color: blue;"><b>schmeckt nicht</b></span>) bis 100 (<b><span style="color: blue;">schmeckt hervorragend</span></b>) bewerten. Im Datensatz werden die durchschnittlichen Bewertungen der unterschiedlichen Probanden für jedes Brot dargestellt.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Außerdem wurde die Härte der Kruste ermittelt. Das passiert indem die Kraft eines Dornes gemessen wurde, die benötigt wird um durch die Kruste zu brechen. Aus diesen Daten können leicht folgende Mittelwerte und Streuungen ermittelt werden:</div>
</div>
<div>
<div style="text-align: justify;">
<br /></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgdRBfjVdXsyZjBsVMXbItDjKeAZS6CY80RNRoLEajdWTnhm3o2LYE-pl0trdcYDBo7T8XmDCJmKivHDHE_ysAG5xBM6YwEONrqvvTjTjytKJqHYWZltlKjql0M5pWZ6P9iegR9JDPhf4w/s1600/mean_sd.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgdRBfjVdXsyZjBsVMXbItDjKeAZS6CY80RNRoLEajdWTnhm3o2LYE-pl0trdcYDBo7T8XmDCJmKivHDHE_ysAG5xBM6YwEONrqvvTjTjytKJqHYWZltlKjql0M5pWZ6P9iegR9JDPhf4w/s1600/mean_sd.jpg" /></a></div>
</div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
<span style="font-family: 'Times New Roman'; text-align: justify;">Ein möglicher Weg die Daten zu visualisieren wäre der Folgende:</span><br />
<span style="font-family: 'Times New Roman'; text-align: justify;"><br /></span>
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi4xxiE1B7I4lK0XvCsKlc4mbrUZzZAJSUqzXdV1hz_-Bx523GPTTHwVicifh7v9s7_jVhwb_zGW-SMerpeREzioiblfQ9KcwaethLL46nHwt7rSJZQZJEKGLRnl-dgfGz-gd1c6dPxAxo/s1600/jmp_kruste.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi4xxiE1B7I4lK0XvCsKlc4mbrUZzZAJSUqzXdV1hz_-Bx523GPTTHwVicifh7v9s7_jVhwb_zGW-SMerpeREzioiblfQ9KcwaethLL46nHwt7rSJZQZJEKGLRnl-dgfGz-gd1c6dPxAxo/s1600/jmp_kruste.jpg" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Balkeniagramm (JMP10 Pro)</span></td></tr>
</tbody></table>
<div style="min-height: 15px;">
<div style="text-align: justify;">
Während das Balkendiagramm für die Kruste sinn macht ist es eher fraglich, ob wie ein vergleichbares Balkendiagramm auch für den Geschmack anwenden sollten. Gehen wir von folgendem Diagramm aus:</div>
<div class="separator" style="clear: both; text-align: center;">
</div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgu8FBjEXFVi53mQuaBrRmEVVpXcfNKdpJF8909iIPQgdcKTggvq78_nJAxmlIUayNVxEbDqbfjK2WnA8-szLf2HMiNYjo2RwhsUE2yLasA3qD6lrdLcCl8zzVDpqc_jDzhHMtlg88E5qw/s1600/NoDifference.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgu8FBjEXFVi53mQuaBrRmEVVpXcfNKdpJF8909iIPQgdcKTggvq78_nJAxmlIUayNVxEbDqbfjK2WnA8-szLf2HMiNYjo2RwhsUE2yLasA3qD6lrdLcCl8zzVDpqc_jDzhHMtlg88E5qw/s1600/NoDifference.png" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Balkendiagramm für Daten ohne sinnvollen Nullpunkt (ggplot2/R)</span></td></tr>
</tbody></table>
<div style="text-align: justify;">
Scheinbar gibt es im Geschmack keinen wirklichen Unterschied zwischen den vier Gruppen. Alle Gruppen scheinen eine Bewertung von etwa 70 Punkten zu bekommen. Der Balken referenziert jetzt allerdings auf den Null-Punkt als Basis, also als schlechtest mögliche Wertung. Ist dieser Ausgangspunkt sinnvoll?</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
<i>Es stellt sich doch die Frage: Kann es überhaupt ein schlechtestes Brot geben? Gibt es ein Brot mit einem Geschmack von 0 Punkten?</i></div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Zwar haben wir bei der Bewertung des Geschmacks willkürliche Grenzen vorgegeben, doch ist kein wirkliches technisch, sinnvolles Maximum oder Minimum für die Bewertung des Geschmacks erkennbar. Es ist also ohne weiteres denkbar die Y-Skala des Balkendiagrammes anzupassen. Eine vernünftige Basis könnte ja überall sein, nicht zwangsläufig an der 0.</div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEikDy-mYqWF6Y5A_-i-t-ldA8tTpi552QkXVtVvNYbTC83YG2GaWow2ccC90R8HjOWaOZEvBHuCtU06QLW72_Va4sx0C845E-PeH1zaJCeUIiDehtxcs5DyYyL7HLuNRPI13_mt_fGnfEU/s1600/NoDifference.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEikDy-mYqWF6Y5A_-i-t-ldA8tTpi552QkXVtVvNYbTC83YG2GaWow2ccC90R8HjOWaOZEvBHuCtU06QLW72_Va4sx0C845E-PeH1zaJCeUIiDehtxcs5DyYyL7HLuNRPI13_mt_fGnfEU/s1600/NoDifference.png" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Balkendiagramm mit willkürlicher Y-Achse (ggplot2/R)</span></td></tr>
</tbody></table>
<div class="separator" style="clear: both; text-align: justify;">
</div>
<div style="text-align: justify;">
Wie man sieht hängt das Erscheinungsbild des Diagramms dramatisch von der Wahl der Y-Achsen-Skala ab. Nimmt man die Basis 0 scheint es keine Unterschiede zwischen den Gruppen zu geben, wählt man statt dessen eine Basis von 67 bemerkt man, dass es dramatische Unterschiede zwischen <b>billig*kurz</b> und <b>billig*lang</b> zu geben scheint.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Wenn eine Grafik so stark von einer willkürlich getroffenen Entscheidung abhängt kann sie nicht zu einer objektiven Datenanalyse geeignet sein.</div>
<div style="text-align: justify;">
<span style="text-align: left;"><br /></span></div>
<h3>
<span style="font-size: small;">Alternative: Streudiagramm mit Fehlerbalken</span></h3>
<span style="font-size: small;">Die Streudiagramme stellen im Gegensatz zu Balkendiagrammen keine explizite Verbindung zu einem Nullpunkt her!</span><br />
<div style="text-align: justify;">
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgIvRJN8T9dKL0Tps9cmx9KN9cay195BexxJJ6H5LsYUurM1QGhyphenhyphenfwUiogy_pRSFF1h4Nv2MthSd5mubB90lLsGjp-yMjhohARX5U-v8h2b_kyPGw5BdprN7Q1NHdDWgc5uq70qvw8q1lE/s1600/means.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgIvRJN8T9dKL0Tps9cmx9KN9cay195BexxJJ6H5LsYUurM1QGhyphenhyphenfwUiogy_pRSFF1h4Nv2MthSd5mubB90lLsGjp-yMjhohARX5U-v8h2b_kyPGw5BdprN7Q1NHdDWgc5uq70qvw8q1lE/s1600/means.png" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Scatterplot mit Mittelwerten und Fehlerbalken (ggplot2/R)</span></td></tr>
</tbody></table>
Dieses Diagramm beinhaltet drei Komponenten:</div>
</div>
<div>
<ol>
<li style="text-align: justify;"><span style="font-family: Times New Roman;">Die tatsächlichen Beobachtungen (schwarze Punkte)</span></li>
<li style="text-align: justify;"><span style="font-family: Times New Roman;">Die Gruppenmittelwerte (weiße Punkte)</span></li>
<li style="text-align: justify;"><span style="font-family: Times New Roman;">Die zu den Gruppenmittelwerten gehörigen Fehlerbalken.</span></li>
</ol>
<span style="font-family: 'Times New Roman'; text-align: justify;">Wenn man die Mittelwerte von Gruppen vergleichen möchte sind Fehlerbalken ein Hilfsmittel um die Streuung der Daten mit zu Berücksichtigen. Fehlerbalken sind vertikale Linien, die ausgehend vom Mittelwert nach oben und unten unterschiedliche Längen haben können, meist aber symmetrisch sind. Die Symmetrie hängt von der Verteilung der Daten ab, die Länge der Fehlerbalken vom gewählten Kennwert.</span><br />
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Für verschiedene Fragestellungen kommen unterschiedliche Fehlerbalken in Frage.<br />
<br /></div>
</div>
<br />
<div class="nobrtable">
<table border="2" bordercolor="#192286" cellpadding="10" cellspacing="0" style="background-color: #dcdce0; border-collapse: collapse; width: 95%px;"><tbody>
<tr style="background-color: #192286; color: white; padding-bottom: 4px; padding-top: 5px;"><th>Fragestellung</th><th>Fehlerbalken</th></tr>
<tr><td>Beschreibung der Verteilung der Daten<br />
(ohne Annahmen)</td><td>Minimum, Maximum</td></tr>
<tr><td>Beschreibung der Verteilung der Daten<br />
(ohne Annahmen)</td><td>5%- und 95%-Percentil</td></tr>
<tr><td>Beschreibung der Verteilung der Daten<br />
(Annahme: Symmetrische Verteilung)</td><td>Standardabweichung</td></tr>
<tr><td>Beschreibung der Verteilung der Mittelwerte<br />
(Annahme: Symmetrische Verteilung)</td><td>Standardfehler</td></tr>
<tr><td>Beschreibung der Verteilung zweier Mittelwerte<br />
(Zentraler Grenzwertsatz)</td><td>Konfidenzintervalle(95% oder 99%)</td></tr>
<tr><td>Induktion über mehrere Mittelwerte</td><td>korrigierte Konfidenzintervalle (z.B. Bonferroni)</td></tr>
<tr><td>Induktion über mehrere Mittelwerte - Varianzanalyse (Annahmen: gleiche Streuung in den Gruppen Normalverteilte Residuen)</td><td>Konfidenzintervalle mit gepoolter Streuung</td></tr>
</tbody></table>
</div>
<br />
<div style="text-align: justify;">
Induktion meint hier, das von der Daten einer Stichprobe ein Schlussfolgerung über die Grundgesamtheit abgeleitet wird. Z.B. weil sich die dargestellten Daten so deutlich unterschiedlich darstellen, wird entschieden das in Zukunft nur noch der bessere Mehltyp verwendet wird. Die alternative Entscheidung wäre würde für dieses Beispiel lauten: Weil sich die Mehltypen nur geringfügig unterscheiden können auch in Zukunft beide Mehlttypen verwendet werden.</div>
<div style="text-align: justify;">
<br />
Für diese Art der Schlüsse, die auf der Basis von Daten abgeleitet werden gelten besondere Regeln, die auch die Art der Fehlerbalken bestimmen.</div>
<div style="text-align: justify;">
<br /></div>
<div>
<div style="text-align: justify;">
Die meisten Balkendiagramme für Mittelwerte, auch wenn Sie Fehlerbalken enthalten definieren die Fehlerbalken nicht, obwohl diese abgebildet werden. Offensichtlich gehen die Ersteller davon aus, dass es nur eine sinnvolle Art von Fehlerbalken gibt. Ohne eine deutliche Angabe über die Art des gewählten Fehlerbalkens ist dieser nicht interpretierbar und daher sinnlos!<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhsQfCWegmquDuMyuNbR-2pVfRhSJlIuhTiU9dYbi0NDzM7UJSXmuE4y01hb0ltb57hEc30DLq9gdkzF2xk2HXHPvhsM9HjdmdQZ7eLQI4wDg8bcrKa-QGkm6929bUCDDc7BKFlvxAQSXo/s1600/means_several.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhsQfCWegmquDuMyuNbR-2pVfRhSJlIuhTiU9dYbi0NDzM7UJSXmuE4y01hb0ltb57hEc30DLq9gdkzF2xk2HXHPvhsM9HjdmdQZ7eLQI4wDg8bcrKa-QGkm6929bUCDDc7BKFlvxAQSXo/s640/means_several.png" height="131" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Übersicht über verschiedene Fehlerbalken (ggplot2/R - Klicken zum Vergrößern)</span></td></tr>
</tbody></table>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhKkHIGOwey1yLY5ueq8KBamhabMZ7tT4DwRd7IlUzQo-RuGG99VtalAxk61rO0LhjGPAq282jWRVLz4MGAS1UIcUlBItsb9Q8BLQxdPxQ-TC2yU5pWOm13E1wPHzNWlD48v8nr8OrtyA4/s1600/jmp_ci.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhKkHIGOwey1yLY5ueq8KBamhabMZ7tT4DwRd7IlUzQo-RuGG99VtalAxk61rO0LhjGPAq282jWRVLz4MGAS1UIcUlBItsb9Q8BLQxdPxQ-TC2yU5pWOm13E1wPHzNWlD48v8nr8OrtyA4/s1600/jmp_ci.jpg" /></a></td></tr>
<tr><td class="tr-caption" style="font-size: 13px;"><span style="font-size: small;">Mittelwerte mit Fehlerbalken (Konfidenzintervall 95% - JMP10 Pro)</span></td></tr>
</tbody></table>
Werden im Rahmen der Interpretation induktive Schlüsse gezogen, wie z.B. das sich die beiden Mehltypen im mittleren Geschmack unterscheiden, so muss der Fehlerbalken mindestens ein Konfidenzintervall abbilden, um diesen Schluss graphisch zu untermauern.<br />
<br />
<h3>
Ausblick</h3>
In einem weiteren Eintrag werden wir auf spezielle Graphen zur Visualisierung von Effekten in verschiedenen Studientypen eingehen. Das beinhaltet die Visualisierung in Mixed Models.</div>
</div>
<div style="min-height: 15px;">
<div style="text-align: justify;">
</div>
</div>
Anonymoushttp://www.blogger.com/profile/10837106230553658640noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-4515376248468822112013-05-23T09:08:00.001+02:002013-05-31T09:35:30.813+02:00Balkendiagramme einige allgemeine Hinweise<span style="text-align: justify;">Im Rahmen der <a href="http://blogs.sas.com/content/sasdach/2013/04/09/leitmotiv-datenvisualisierung-ihre-meinung-zahlt/">Blogparade</a> von SAS zum Thema Datenvisualisierung gibt Bertram Schäfer (CEO der Firma Statcon) Hinweise zur Verwendung und Anpassung von Balkendiagrammen.</span><br />
<span style="text-align: justify;"><br /></span>
<a name='more'></a><br />
<h3>
<span style="font-size: small;">Does und Dont’s</span></h3>
<div style="text-align: justify;">
Ein
Balken stellt faktische eine Verbindung zwischen einem Wert auf der
Y-Achse und dem Null-Punkt auf der Y-Achse dar. Es gibt zahlreiche
Beispiele in denen der Null-Punkt nicht abgebildet wird, bzw. keinen
inhaltlichen Sinn hat. Dann sollte man auf die Anwendung von
Balkendiagrammen verzichten. Stattdessen wären Punkt- oder
Liniendiagramme das geeignetere Werkzeug.<br />
<br />
Schönste Beispiel dazu gibt es unter anderem im <a href="http://www.bildblog.de/tag/diagramme/">Bildblog</a>.<br />
<h4>
3D und 2D</h4>
</div>
<div style="text-align: justify;">
Gerade in den
Medien werden immer wieder 3D-Balkendiagramme verwendet. Was auf den
ersten Blick „hübscher“ aussehen mag, verbirgt auf den zweiten Blick
wichtige Information.</div>
<div style="text-align: justify;">
<br /></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFhMKtqYfzH901eNcNoe6iuW-f63kEptL4NIasISGEWERxHXS8T40IJXAUlmvZx9ktOOoGlSjTXjfMDKS04PCR09HHlPW3qVBpMfiE7XZCk4zIql0LxrYRpGnPpBA0PNcKbtfjTm7XWWg/s1600/3DBarChart.jpg" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFhMKtqYfzH901eNcNoe6iuW-f63kEptL4NIasISGEWERxHXS8T40IJXAUlmvZx9ktOOoGlSjTXjfMDKS04PCR09HHlPW3qVBpMfiE7XZCk4zIql0LxrYRpGnPpBA0PNcKbtfjTm7XWWg/s1600/3DBarChart.jpg" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-size: small;">3D-Balkendiagramm (Excel)</span></td></tr>
</tbody></table>
<div style="text-align: justify;">
Wie hoch sind in
diesem Balkendiagramm wohl die einzelnen Balken? Wenn es um die
Vermittlung von Wissen geht eignet sich die folgende Variante wesentlich
besser.<br />
<br /></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjmEAu7y3TgiuOYMGJkBX2vrEMevv8cxpfl6B0XnTviShMTHWrlpbJHS-tVl01pR6E6T8GfO_iRjSfe7rMoaH49wqYCKLZ79AnkPIWv_LV__99k3LTUHeQB0IoDAhR9pjd6wGb7VPtCaOk/s1600/jmp.PNG" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjmEAu7y3TgiuOYMGJkBX2vrEMevv8cxpfl6B0XnTviShMTHWrlpbJHS-tVl01pR6E6T8GfO_iRjSfe7rMoaH49wqYCKLZ79AnkPIWv_LV__99k3LTUHeQB0IoDAhR9pjd6wGb7VPtCaOk/s1600/jmp.PNG" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-size: small;">Balkendiagramm (JMP 10)</span></td></tr>
</tbody></table>
<div style="min-height: 14px; text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Grundsätzlich
sind die Höhenunterschiede der Balken in der pseudo-dreidimensionalen
Darstellung am Bildschirm oder auf Papier mit wenigen Ausnahmen,
schwerer für das menschliche Auge zu erfassen als in der
zweidimensionalen Darstellung.</div>
<div>
Große Balkendiagramme</div>
<div>
In
praktischen Anwendungen begegnet man oft vielen Kategorien und somit
vielen Balken. Verschaffen wir uns einen Überblick über die letzten
Bahnfahrten eines meiner Mitarbeiter.</div>
<div>
<br /></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgqk74F-pTvYIw6CzjzT8yro3FtiKBMP_kuoblslAAA_2zsnc9PlKkLwsWGtLd7Ym3a85j1McnJzd5x22aJ4ohxFeUla2wgpcjQhR5Czl2M0dwFM1VveLGZilr-CWAFgdMU_hQ4mpVPYJQ/s1600/LongBarChart.png" style="margin-left: auto; margin-right: auto;"><img border="0" height="212" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgqk74F-pTvYIw6CzjzT8yro3FtiKBMP_kuoblslAAA_2zsnc9PlKkLwsWGtLd7Ym3a85j1McnJzd5x22aJ4ohxFeUla2wgpcjQhR5Czl2M0dwFM1VveLGZilr-CWAFgdMU_hQ4mpVPYJQ/s640/LongBarChart.png" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-size: small;">2D-Balkendiagramm (ggplot2/R)</span></td></tr>
</tbody></table>
<div>
Was fällt auf?</div>
<ul>
<li style="margin: 0px;">Ganz besonders wenn die Anzahl der Balken groß
wird, sollte es dem Leser so einfach wie möglich gemacht werden die
Information der Grafik zu erfassen.</li>
<li style="margin: 0px;">Die Beschriftung der Balken sollte lesbar sein, ohne dass man den Kopf drehen muss.</li>
<li style="margin: 0px;">Die Anzahl der Markierungen auf der Y-Achse
sollte so gewählt sein, dass sie ein Ablesen von Datenwerten
unterstützt. D.h. Sie sollten nicht zu viele und nicht zu wenige
„Tick-Marks“ verwenden. Faustregel: Nicht weniger als 3 nicht mehr als
10.</li>
<li style="margin: 0px;">Die Erscheinung des Diagramms hängt stark
Höhen-Breiten-Verhältnis der Grafik ab. Wählen Sie wenn möglich ein
Verhältnis von etwa 1.6181:1 („goldener Schnitt“) oder 1.85:1 bzw. 2.5:1
für Diagramme mit vielen Balken (oft: jährliche/monatliche Daten über
einen langen Zeitraum).</li>
</ul>
<div>
Oft hilft es einem Balkendiagramm, wenn man es im Winkel
von 90° rotiert. Gerade längliche Beschriftungen sind dann oft besser
lesbar, als in der „natürlichen“, horizontalen Ausrichtung.</div>
<div>
<br /></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjfyXD0lNQM1fJ8eJTZqSCwH7Uw3TWXHgYjlODUKn7XnJh5fqfXZAzaoKxxNpWLErPKuhfelWe3Ipgm9GSsUiKJVqpai_7GcTxcGf3U8xpz_IcyVPOQwH7A-nv9t9js_W2p0OtFUOFfWHc/s1600/VerticalBarChart.png" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjfyXD0lNQM1fJ8eJTZqSCwH7Uw3TWXHgYjlODUKn7XnJh5fqfXZAzaoKxxNpWLErPKuhfelWe3Ipgm9GSsUiKJVqpai_7GcTxcGf3U8xpz_IcyVPOQwH7A-nv9t9js_W2p0OtFUOFfWHc/s1600/VerticalBarChart.png" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;"><span style="font-size: small;">Vertikal orientiertes 2D-Balkendiagramm (ggplot2/R)</span></td></tr>
</tbody></table>
Anonymoushttp://www.blogger.com/profile/10837106230553658640noreply@blogger.com2tag:blogger.com,1999:blog-276389194507638774.post-33924839806590995272013-05-15T12:00:00.000+02:002013-05-15T12:00:04.575+02:00Ingenieursmathematik für Jedermann<div style="text-align: justify;">
Unser geschätzter Kollege Prof. Dr. Dr. David Meintrup bewirbt gerade seinen Online-Kurs zum Thema "Ingenieursmathematik für Jedermann" im Rahmen der MOOC Production Fellowship. Hoffentlich kann das folgende Video, Sie dazu <a href="https://moocfellowship.org/submissions/ingenieurmathematik-fur-jedermann">hier</a> für Ihn abzustimmen.
<br />
<br />
<center>
<iframe allowfullscreen="" frameborder="0" height="315" src="http://www.youtube.com/embed/_agXfmFAtrQ" width="560"></iframe></center>
<center>
<br /></center>
<h4>
<center style="text-align: left;">
<a name='more'></a><br /></center>
<center style="text-align: left;">
Hintergrund (<a href="https://moocfellowship.org/info">Quelle</a>)</center>
</h4>
<center style="text-align: left;">
<div style="background-color: white; font-family: Ubuntu, sans-serif; font-size: 14px; line-height: 23px; margin-bottom: 10px; text-align: -webkit-auto;">
Weltweit begreifen viele Hochschulen die Digitalisierung als Möglichkeit für Innovationen in der Hochschullehre. Mit dem Wettbewerb „MOOC Production Fellowship“ greifen der Stifterverband für die Deutsche Wissenschaft und iversity diese Entwicklung auf.</div>
<div style="background-color: white; font-family: Ubuntu, sans-serif; font-size: 14px; line-height: 23px; margin-bottom: 10px; text-align: -webkit-auto;">
Ziel des Wettbewerbs ist es, die Entwicklung innovativer Konzepte für Massive Open Online Kurse (kurz MOOC) anzustoßen und die Umsetzung von zehn Kurskonzepten zu ermöglichen. Der Stifterverband und iversity hoffen, durch diesen Wettbewerb das große Potential, das in einer Nutzung der digitalen Möglichkeiten steckt, öffentlichkeitswirksam aufzuzeigen und so der Organisationsentwicklung der Hochschulen in Zeiten des digitalen Wandels einen wichtigen Impuls zu geben.</div>
<div style="background-color: white; font-family: Ubuntu, sans-serif; font-size: 14px; line-height: 23px; margin-bottom: 10px; text-align: -webkit-auto;">
Prämiert werden bis zu zehn Lehrende oder Lehrteams für ihr Konzept eines innovativen Massive Open Online Course (MOOC). Die Preisträger* erhalten je 25.000 Euro Förderung zur Realisierung ihres Online-Kurses. Ziel ist es, im Rahmen des Wettbewerbs fünf Kurse für das WS 2013/14 und weitere fünf Kurse für das SS 2014 zu produzieren. Die Online-Kurse werden auf der von iversity entwickelten Plattform kostenfrei angeboten und stehen allen Interessierten offen.</div>
</center>
<center style="text-align: left;">
<br /></center>
</div>
Sebastian Hoffmeisterhttp://www.blogger.com/profile/14491012569433491752noreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-22394961621978524592013-02-14T15:24:00.000+01:002013-02-14T08:51:13.488+01:00Spielen mit JMP<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjsnVpJnBZUPWnzR0txIDVrmfa2Wc_aWdalQlMQuhPY_InKU4ru8TsR-1g6-pRK3t-HBLBe-jMTn1Yr9duVVt2zO_GI_rWb92eo79XxyYTh_BLJg4tehRVMR_IjFMxIirbrtUghyEHdbQg/s1600/Logo.png" imageanchor="1" style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><img border="0" height="220" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjsnVpJnBZUPWnzR0txIDVrmfa2Wc_aWdalQlMQuhPY_InKU4ru8TsR-1g6-pRK3t-HBLBe-jMTn1Yr9duVVt2zO_GI_rWb92eo79XxyYTh_BLJg4tehRVMR_IjFMxIirbrtUghyEHdbQg/s1600/Logo.png" width="640" /></a></div>
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<b>14.02.2013 - Ein kleines Update:</b><br />
<br />
Die Kollegen von SAS fanden diesen Eintrag interessant und haben ihn auch auf ihrem <a href="http://blogs.sas.com/content/jmp/">JMP-Blog</a> veröffentlicht (hier der direkte <a href="http://blogs.sas.com/content/jmp/2013/02/13/test-your-geographical-knowledge-with-jmp/">Link</a>).<br />
<br />
Vielen Dank dafür! Ich wünsche allen viel Spass mit dem Spiel!<br />
Sebastian Hoffmeister<br />
<br />
<br />
<a name='more'></a><br />
<br />
<br />
Das Spiel ist einfach:<br />
1. Öffnen Sie die folgende JMP-Datei: <a href="http://www.statcon.de/blog/Capitals.jmp">JMP-File</a><br />
2. Starten Sie das angehängte Script:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg1lTS1JAaMebJz_v_dPx4Dx-34_C5f2AOijg_In6EFHf9UjavunnZfsXX9cUDJSCz_WFFt7PvD4X9d94WWF91lCrx-ZIjaMEzOKo5hAg86N-fCZpbsES2Gjlz9HU8jjJcvFuY5Xhkg_cs/s1600/script_highlighted.png" imageanchor="1" style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><img border="0" height="212" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg1lTS1JAaMebJz_v_dPx4Dx-34_C5f2AOijg_In6EFHf9UjavunnZfsXX9cUDJSCz_WFFt7PvD4X9d94WWF91lCrx-ZIjaMEzOKo5hAg86N-fCZpbsES2Gjlz9HU8jjJcvFuY5Xhkg_cs/s1600/script_highlighted.png" width="640" /></a></div>
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
3. Klicken Sie mit der Maus auf die Weltkarte an der Stelle an der Sie den oben abgefragten Ort vermuten. Sie haben 60 Sekunden Zeit möglichst viele Orte zu erraten.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg_4vI4leTlZC1oyZihRrtSedkma0jzZIob34F0Ln73jrTFQmdoLCU83J-K3Hm3NKlXAFDyIxbZ5ZDaoH9igwn62GGEuJp6EKwWAjSUXdGEW3VhjbGZPWrAQPH3LGER1YpW_6Mdl6jJNOs/s1600/map_highlighted.png" imageanchor="1" style="clear: left; float: left; margin-bottom: 1em; margin-right: 1em;"><img border="0" height="398" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg_4vI4leTlZC1oyZihRrtSedkma0jzZIob34F0Ln73jrTFQmdoLCU83J-K3Hm3NKlXAFDyIxbZ5ZDaoH9igwn62GGEuJp6EKwWAjSUXdGEW3VhjbGZPWrAQPH3LGER1YpW_6Mdl6jJNOs/s1600/map_highlighted.png" width="640" /></a></div>
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
4. Posten Sie Ihren Highscore in den Kommentaren. Es warten Ruhm und Ehre!<br />
<br />
Frohe Weihnachtstage & einen guten Rutsch ins neue Jahr wünscht das Statcon-Team!<br />
<br />Unknownnoreply@blogger.com5tag:blogger.com,1999:blog-276389194507638774.post-40581430519762027422013-01-09T15:33:00.000+01:002013-01-11T12:28:41.059+01:00Why we need the lower-order-terms ...This is the second article in our series about <a href="http://statistiksoftware.blogspot.de/2013/01/a-discussion-on-non-hierarchical.html">non-hierarchical models</a>. Here i want to point out the problem of non-hierarchical regression models meaning models that include some higher-order-terms without the corresponding lower-order-terms.
<br />
<div style="text-align: -webkit-auto;">
<br /></div>
<div style="margin-left: 1em; margin-right: 1em;">
<br /></div>
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpcyU5ehAjcg4F1sYpPLy_sCX85JIvyAxUI0GEuDXUDQjy9jhgE4MVSYm9x1ZoS8yeRE4_KSy3sg9El9lEnPFRj0h2k1Epb-FWF4tX2I3_RhkiZddRhENI6LNbK22Aap4wyVi1uXxLi6Q/s1600/Reihenfolge3.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="118" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhpcyU5ehAjcg4F1sYpPLy_sCX85JIvyAxUI0GEuDXUDQjy9jhgE4MVSYm9x1ZoS8yeRE4_KSy3sg9El9lEnPFRj0h2k1Epb-FWF4tX2I3_RhkiZddRhENI6LNbK22Aap4wyVi1uXxLi6Q/s400/Reihenfolge3.png" width="400" /></a></div>
<br />
<br />
<a name='more'></a><br />
<br />
Have you ever received such a message when fitting a linear model?<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEivjY2rQhNiG7ULJBqFPWlyzqaugg0dUyMIlOa01N8EhGrAS7hq_aWhmfRemIeVPIesje6UFFyXHTb-Y2B833QfI_z2WhQHhUGS0HrF-mJeEX2mkmsmxF464-cB1EgAJn7Z8rJOxTpxphU/s1600/dx.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEivjY2rQhNiG7ULJBqFPWlyzqaugg0dUyMIlOa01N8EhGrAS7hq_aWhmfRemIeVPIesje6UFFyXHTb-Y2B833QfI_z2WhQHhUGS0HrF-mJeEX2mkmsmxF464-cB1EgAJn7Z8rJOxTpxphU/s1600/dx.png" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Design Expert - Warning</td></tr>
</tbody></table>
Or this?<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi_zkahxx7xssTbXgWraItm-TMbm0d7_Fe8vczkb6uNq2vKA4lyY66evH0Mt9wmwK58inc3H89b5i2hxFvw2GLdlj_cgwbm3_N-Mfl53DrasCUuzrR4F082qPNPkup590sAhAt-tK4YcWs/s1600/jmp.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi_zkahxx7xssTbXgWraItm-TMbm0d7_Fe8vczkb6uNq2vKA4lyY66evH0Mt9wmwK58inc3H89b5i2hxFvw2GLdlj_cgwbm3_N-Mfl53DrasCUuzrR4F082qPNPkup590sAhAt-tK4YcWs/s1600/jmp.png" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">JMP Warning</td></tr>
</tbody></table>
When you face these warnings you probably tried to fit a model containing higher-order-terms without using the corresponding main effects. Typical examples are models containing two-factor-interactions ($x_1*x_2$) without using the main effects in the model ($x_1$ and $x_2$ itself). The same applies to polynomial models containing higher order effects (quadratic, cubic, ...) leaving out the linear term.<br />
<br />
So what is the problem of non-hierarchical models? Why these warnings? Why do we usually recommend to respect model-hierarchy? The main argument is the following:<br />
<br />
<b>Non-hierarchical models aren't invariant versus location shifts</b><br />
<br />
Assume we are using a non-hierarchical model like $taste_i = \beta_0 + \beta_1 temp_i*time_i$. We have a response $taste$ and two predictors $temp$ and $time$. The model only uses the interaction of both predictors to explain the response.<br />
<table align="center" style="text-align: right;">
<tbody>
<tr>
<td><b>temp<b></b></b></td><td><b>time<b></b></b></td><td><b>taste<b></b></b></td><td><b>ctemp</b><td/><td><b>ctime</b><td/>
</tr>
<tr>
</tr>
<tr>
<td>°C</td><td>min </td><td>--- </td><td>°C<td/><td>min<td/>
</tr>
<tr>
</tr>
<tr>
<td>190</td><td>10</td><td>2</td><td>5.8<td/><td>-11.7<td/>
</tr>
<tr>
</tr>
<tr>
<td>195</td><td>15</td><td>5</td><td>10.8<td/><td>-6.7<td/>
</tr>
<tr>
</tr>
<tr>
<td>200</td><td>20</td><td>3</td><td>15.8<td/><td>-1.7<td/>
</tr>
<tr>
</tr>
<tr>
<td>175</td><td>30</td><td>8</td><td>-9.2<td/><td>8.3<td/>
</tr>
<tr>
</tr>
<tr>
<td>180</td><td>30</td><td>6</td><td>-4.2<td/><td>8.3<td/>
</tr>
<tr>
</tr>
<tr>
<td>165</td><td>25</td><td>4</td><td>-19.2<td/><td>3.3<td/>
</tr>
<tr>
</tr>
</tbody></table>
<br />
Lets start with analysing the full model: $taste_i = b_0 + b_1 \cdot temp_i + b_2 \cdot time_i + b_{12} \cdot temp_i\cdot time_i + \epsilon_i$<br />
<br />
The estimated model is:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjLQIxrAMeb6CjXz5NHWmj_xhXYbnwOpu37S7msxLgT5isNFTqSlbVbWihXfXlw4aBzyZnH8EHyQzsmPtTlAdWDN7OSgFfdAU3Ct36xmdfeh0ymuAgDZA_Rv1nGznOedizQ6tdqF7Z-iHw/s1600/full+model.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjLQIxrAMeb6CjXz5NHWmj_xhXYbnwOpu37S7msxLgT5isNFTqSlbVbWihXfXlw4aBzyZnH8EHyQzsmPtTlAdWDN7OSgFfdAU3Ct36xmdfeh0ymuAgDZA_Rv1nGznOedizQ6tdqF7Z-iHw/s1600/full+model.jpg" /></a></div>
<br />
But the vifs seem to be rather problematic:<br />
<br />
<table align="center" style="text-align: right;">
<tbody>
<tr>
<td></td><td></td><td><b> VIF</b></td><td></td><td><b>VIF centered</b></td>
</tr>
<tr>
</tr>
<tr>
<td><b> temp<b> </b></b></td><td></td><td>358 </td><td></td><td>2.54</td></tr>
<tr>
</tr>
<tr>
<td><b> time<b> </b></b></td><td></td><td>8990 </td><td></td><td>2.75</td></tr>
<tr>
</tr>
<tr>
<td><b> temp*time<b> </b></b></td><td></td><td>7074 </td><td></td><td>1.72</td></tr>
<tr>
</tr>
</tbody></table>
<br />
To avoid the problem of multicollinearity let us center the factors. Then recalculate the model:<br />
<div>
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiqhRES5vRW7M9hIGYkbDpJnblLVWEaFOOArimNv_upVm6P6dw0GmzZMzRa5vd1aO3wIPuBQ4URMFy1RmhJRa57k09KC2YSZdkfCxr6zZa2bCwISijJuv6ok-Zhab4n-OszgZOWjgYwAx8/s1600/cfull+model.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiqhRES5vRW7M9hIGYkbDpJnblLVWEaFOOArimNv_upVm6P6dw0GmzZMzRa5vd1aO3wIPuBQ4URMFy1RmhJRa57k09KC2YSZdkfCxr6zZa2bCwISijJuv6ok-Zhab4n-OszgZOWjgYwAx8/s1600/cfull+model.jpg" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
<div>
Of course the main effect estimates change but the estimate and p-value of the interaction are still the same! The VIFs are all below 3 now.</div>
<div>
<br /></div>
<div>
<b>Results</b></div>
<div>
<ul>
<li>Use centered data as it removes a multicollinearity problem.</li>
<li>The 2-factor-interaction is significant.</li>
<li>The centered temperature factor is not significant.</li>
</ul>
</div>
<div>
The question is: May we remove the main-effect temperature now? Go one step further: Let us figure out what happens if we remove both main effects.</div>
<div>
<ol>
<li>As we are using only one factor now we do not care for multicollinearity any more. So we might use the original data.</li>
<li>Estimate the pure interaction model: $taste_i = b_0 + b_{12} temp_i \cdot time_i + \epsilon_i$.</li>
</ol>
</div>
<div>
<br />
The estimated <b>model</b> is the following:<br />
<div>
</div>
<table>
</table>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhTqaVnwk8lyf-xza7tMbR3ZJRZTXO008nBnESMN90K8TbIZGt9iZueJufOc4NWle3ZZwPtDYpzUPPXwL-R-L5kgKP1WXTTsKR95ET2apA8-dDKVgwQgiOXkV2IbHdjCqDKc39BlANVG-U/s1600/model1.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhTqaVnwk8lyf-xza7tMbR3ZJRZTXO008nBnESMN90K8TbIZGt9iZueJufOc4NWle3ZZwPtDYpzUPPXwL-R-L5kgKP1WXTTsKR95ET2apA8-dDKVgwQgiOXkV2IbHdjCqDKc39BlANVG-U/s1600/model1.jpg" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Non-hierarchical Model (computed with R)</td></tr>
</tbody></table>
<div>
<span style="text-align: center;">The two-factor-interaction seems to be significant at a level of significance of 0.1 (typical level of significance in a screening situation).</span><br />
<span style="text-align: center;"><br /></span>
<br />
Finally estimate the <b>same model for the centered data</b> (just for comparison):<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9TqEJM_GiHu7VHwZNe7XwLKhtV5GyRHPLTmMx1WIPhLFlNxvNEj2y_AUXKZCOOTrXVuGY0kvsfRA01ms5IIV563GYr_Uqjem-Eia_wDsIfRUFqbv88NoWOoTHW717FdRS8t0IAJspLfw/s1600/cimodel.jpg" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9TqEJM_GiHu7VHwZNe7XwLKhtV5GyRHPLTmMx1WIPhLFlNxvNEj2y_AUXKZCOOTrXVuGY0kvsfRA01ms5IIV563GYr_Uqjem-Eia_wDsIfRUFqbv88NoWOoTHW717FdRS8t0IAJspLfw/s1600/cimodel.jpg" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Non-hierarchical Model for centered data (in R)</td></tr>
</tbody></table>
<div class="separator" style="clear: both; text-align: center;">
</div>
<span style="text-align: center;">As you can see the results change heavily. While we did not touch the relation between $temp,time$ and $taste$ (we only subtracted the mean of each variable) the interaction in the first model is significant it is not in the second.</span><br />
<span style="text-align: center;"><br /></span>
<span style="text-align: center;">What happened here and why might this be a problem?</span><br />
<br />
<span style="text-align: center;"><b>Mathematical motivation</b></span><br />
<br style="text-align: center;" />
<span style="text-align: center;">If you are using an interaction-model without the main effects, the model is not invariant to location shifts of the factors. If the factors in the interaction model $y_i = \beta_0 + \beta_1 x_i*z_i$ are centered the model is extended to some kind of full model, as: $$ y_i = \beta_0 + \beta_1 (x_i - \bar{x})*(z_i - \bar{z}) = $$</span><br />
<span style="text-align: center;">$$ y_i = \beta_0 + \beta_1 (x_i*z_i - \bar{x}*z_i - x_i*\bar{z} + \bar{x}\bar{z})$$</span><br />
<span style="text-align: center;">You see, that this new model contains the pure main effects in the terms $z_i*\bar{x}$ and $x_i*\bar{z}$ as $\bar{x}$ and $\bar{z}$ are only constants.</span><br />
<br style="text-align: center;" />
<span style="text-align: center;"><b>Why is this bad?</b></span><br />
<br style="text-align: center;" />
<span style="text-align: center;">Two simple arguments:</span><br />
<br style="text-align: center;" />
<span style="text-align: center;">1. Especially in the presence of quadratic effects we often want to center variables to avoid multi-colinearity-problems. So we will often be in the situation that this problem occurs.</span><br />
<span style="text-align: center;">2. Inference should be independent from units. Think of a temperature as a predictor: There shouldn't be a difference in the models if you change degree Celsius to degree Kelvin. But this is exactly what happens in non-hierarchical models.</span><br />
<br style="text-align: center;" />
<span style="text-align: center;"><b>Literature</b></span><br />
<span style="text-align: center;">[1.] Discussion on <a href="http://stats.stackexchange.com/">CrossValidated</a>: <a href="http://stats.stackexchange.com/questions/27724/do-all-interactions-terms-need-their-individual-terms-in-regression-model">Link</a></span></div>
<br />
<br /></div>
</div>Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-82310151629938846492013-01-08T15:39:00.000+01:002013-01-09T11:38:05.666+01:00Why we need the intercept ...<div>
<span style="text-align: left;">This is the first article in the series on </span><a href="http://statistiksoftware.blogspot.de/2013/01/a-discussion-on-non-hierarchical.html" style="text-align: left;">non-hierarchical regression models</a><span style="text-align: left;">. In this article we discuss the pitfalls when performing a regression without using the intercept term.</span></div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjyTdI1acUjEYCFs1sASfpQZoFk2i49ZseI-eSsCJMFWQ87rLC6DEuT6qPMu9p5LDU0ZTw4xiy37tKJvxUkHg3sQn4uvvgCRvFSbIFDoZRPVfajqU8h52wSvvp389cCs1UpcsHMz5OXgVg/s1600/Reihenfolge2.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="118" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjyTdI1acUjEYCFs1sASfpQZoFk2i49ZseI-eSsCJMFWQ87rLC6DEuT6qPMu9p5LDU0ZTw4xiy37tKJvxUkHg3sQn4uvvgCRvFSbIFDoZRPVfajqU8h52wSvvp389cCs1UpcsHMz5OXgVg/s400/Reihenfolge2.png" width="400" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
A common question is, if the intercept-term may be removed from a regression analysis in case it is not significant. Most of the time the answer to this question should be "No!" and I totally agree with that answer. But of course we would like to understand why!<br />
<br />
<a name='more'></a>We will use the following terminology:<br />
<br />
1. A <b>full model</b> is a model containing the intercept and main effects: $y = b_0 + b_1\cdot x + \epsilon$<br />
2. A <b>no-intercept-model</b> reduces the <b>full model</b> by its intercept: $y = b_1 \cdot x + \epsilon$<br />
3. The <b>intercept-only-model</b> (often: baseline model) uses only the intercept to explain the data: $y = b_1 + \epsilon$<br />
<br />
<h3>
<b>Problems of no-intercept-models</b></h3>
<ol>
<li>$R^2$ is not useful any more.</li>
<li>The slope estimators might be biased.</li>
</ol>
<h3>
<b>Explanation</b></h3>
The first commonly mentioned problem (see [1.]) is that the $R^2$-statistic is not useful anymore if the intercept is not included in a linear regression model. Usually $R^2$ is interpreted as the amount of variation that is explained by the model.<br />
<br />
$$R^2 = \frac{Model SS}{Total SS} = 1 - \frac{Residual SS}{Total SS}$$<br />
<br />
More precise: The <b>full model</b> is compared to a reduced model, which in this case is the <b>intercept-only-model</b>. So what do we do for an <b>no-intercept-model</b>? Well we will not compare it to the <b>intercept-only-model</b>. Both models are completely independent from each other so there is no sense in comparing. Instead most software packages (R, JMP, SAS, DX, SPSS, …) compare the model without intercept to a reference model that has lower order. That is a model with no intercept and no other effects.<br />
<br />
$$\text{Noise-Model}: y = \epsilon$$<br />
<br />
One might call it a noise-model. Of course this is no real model (it does not explain anything) and any comparison with it is not very useful. Actually there is no real reference model we could use to compare our <b>no-intercept-model</b> with. So there is no interpretable $R^2$ for models w/o intercept.<br />
<br />
None the less statistics software will present an $R^2$ for <b>no-intercept models</b>. But as the interpretation of the $R^2$ is lost we cannot use it to evaluate the model quality. One can even show that the $R^2$ of a<b> no-intercept model</b> will usually be higher compared to the $R^2$ of a full model (see the mathematical details for that).<br />
<br />
For the example presented in the graph below the $R^2s$ (calculated with R) are:<br />
<br />
<table style="text-align: right;" align="center">
<tbody>
<tr>
<td></td> <td><b>$R^2$</b></td>
</tr>
<tr>
<td><b>Full Model</b></td> <td>0.7846</td>
</tr>
<tr>
<td><b>No-Intercept-Model</b></td> <td>0.9114</td>
</tr>
</tbody></table>
<br /><div>
It is obvious that this is not a reasonable result. The red line is clearly <b>not</b> the better model!</div>
<div>
<br /></div>
<br />
The second problem that arises is that the least squares estimator for the slopes in a no-intercept model are biased (systematically shifted towards larger or smaller values).<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhj8BlQRgA8A0YmiL7w1EJD6GsCbTUB06vFni1cGKiZI25ACYHctf7KO5R0dBwhoeV-VjJFs6Rka2qHYoD5YZkgYa_vrS_nJ7-IlreVrGvssjMzhbex2iOwc4E4dpoyRLieNFpNGbgfwtE/s1600/Origin.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="293" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhj8BlQRgA8A0YmiL7w1EJD6GsCbTUB06vFni1cGKiZI25ACYHctf7KO5R0dBwhoeV-VjJFs6Rka2qHYoD5YZkgYa_vrS_nJ7-IlreVrGvssjMzhbex2iOwc4E4dpoyRLieNFpNGbgfwtE/s400/Origin.png" width="400" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Usual (teal) and regression without intercept (red)</td></tr>
</tbody></table>
<br />
With removing the intercept from the model we impose a restriction so that the regression line goes through the origin (x=0;y=0). The graph shows what happens to the regression line. The blue line is the common regression line the red line is the no-intercept-regression-line. It is heavily pulled down because it has to go through $(0;0)$.<br />
<br />
<h3>
Mathematical details</h3>
<script language="javascript">
function toggle() {
var ele = document.getElementById("toggleText");
var text = document.getElementById("displayText");
if(ele.style.display == "block") {
ele.style.display = "none";
text.innerHTML = "R^2 does not work for no-intercept-models";
}
else {
ele.style.display = "block";
text.innerHTML = "R^2 does not work for no-intercept-models";
}
}
</script>
<img src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEipoDAEakCj9H8mc4SeqD3MKWx-tV610LKIfg833FQ_WOzOrXTJSAdlkMMamXnPe2s-ukzBG8NCr4n1LJa_yNnCzxRPHupssyuvOW2YNSnSW7yYTK6IPxaluddhSCevy6o5typWb9QKJjU/s1600/dropdown.png" style="text-align: center;" />
<a href="javascript:toggle();" id="displayText" style="text-align: center;">R^2 does not work for no-intercept-models</a>
<br />
<br />
<div id="toggleText" style="background-color: gainsboro; border: thin solid Gray; display: none; margin-left: 20px; margin-right: 20px; padding: 20px;">
<br />
So what exactly happens to $R^2$? As noted above the typical $R^2$ is defined as:<br />
<br />
$$ R^2 = \frac{\sum_i (\hat y_i - \bar y)^2 }{ \sum_i (y_i - \bar y)^2} = 1 - \frac{\sum_i (y_i - \hat y_i)^2}{\sum_i (y_i - \bar y)^2}$$<br />
<br />
Formulate the $R^2_0$ for a no-intercept-model $y = b\cdot x$:<br />
<br />
$$ R^2_0 = \frac{\sum_i \hat y_i^2}{\sum_i y_i^2} = 1 - \frac{\sum_i (y_i - \hat y_i)^2}{\sum_i y_i^2}$$<br />
<br />
At the end of the last equation you see that we compare the Residual-Sum-of-Squares with the sum-of-squares of the actual observations. So this is a ratio of the variation of the data around the model (Residual sum of squares) and the magnitude of the response.<br />
<br />
Apart from not really being interpretable $R^2_0$ is often larger than $R^2$. This often leads to the wrong (!) assumption that the model w/o intercept is better in explaining the data than the full model. So how can $R^2_0$ be greater thant $R^2$? Usually we expect $R^2$ to become greater whenever we add more parameters to the model. Now we reduce the model and $R^2$ raises?<br />
<br />
Use $\tilde{y_i}$ for the fitted values of the <b>no-intercept model</b> and $\hat y_i$ for the fitted values of the <b>full model</b>. Then $R_0$ is greater than $R^2$ whenever:<br />
<br />
$$ R^2_0 = 1 - \frac{\sum_i (y_i - \tilde{y}_i)^2}{\sum_i y_i^2} > R^2 = 1 - \frac{\sum_i (y_i - \hat y_i)^2}{\sum_i (y_i - \bar y)^2}$$<br />
$$\Rightarrow \frac{\sum_i (y_i - \tilde{y}_i)^2}{\sum_i y_i^2} < \frac{\sum_i (y_i - \hat y_i)^2}{\sum_i (y_i - \bar y)^2}$$<br />
$$ \Rightarrow \frac{||y-\tilde y||_2^2}{||y-\hat y||_2^2} < \frac{||y||_2^2}{||y-\bar y||_2^2}$$<br />
<br />
Now use $||y||_2^2= ||y - \bar y + \bar y||_2^2 = ||y - \bar y||_2^2 +n\bar y^2$. Then:<br />
<br />
$$ \frac{||y-\tilde y||_2^2}{||y-\hat y||_2^2} < \frac{||y-\bar y||_2^2 + n\bar y^2}{||y-\bar y||_2^2} $$<br />
<br />
$$ \frac{||y-\tilde y||_2^2}{||y-\hat y||_2^2} < 1 + \frac{\bar y^2}{\frac{1}{n}||y-\bar y||_2^2}$$<br />
<br />
The left hand side is allways greater than 1, as the fitted values of the <b>no-intercept model</b> $\tilde{y}$ will always be worse than the fitted values of the <b>full model</b> $\hat y$. The last term of the right hand side is large if the squared mean response is greater than the variance of the response. So $R_0^2$ will be larger than $R^2$ whenever the mean of the response $\bar y$ is much larger than the standard deviation of the response $\sqrt{\frac{1}{n} \sum_i (y_i - y)^2 }$ (forget about the $\frac{1}{n-1}$ for being unbiased :-)).<br />
<br /></div>
<script language="javascript">
function toggle2() {
var ele = document.getElementById("toggleText2");
var text = document.getElementById("displayText2");
if(ele.style.display == "block") {
ele.style.display = "none";
text.innerHTML = "Proof that LS-estimator is biased in no-intercept models";
}
else {
ele.style.display = "block";
text.innerHTML = "Proof that LS-estimator is biased in no-intercept models";
}
}
</script>
<img src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEipoDAEakCj9H8mc4SeqD3MKWx-tV610LKIfg833FQ_WOzOrXTJSAdlkMMamXnPe2s-ukzBG8NCr4n1LJa_yNnCzxRPHupssyuvOW2YNSnSW7yYTK6IPxaluddhSCevy6o5typWb9QKJjU/s1600/dropdown.png" />
<a href="javascript:toggle2();" id="displayText2"> Proof that LS-estimator is biased in no-intercept models</a>
<br />
<div id="toggleText2" style="background-color: gainsboro; border: thin solid Gray; display: none; margin-left: 20px; margin-right: 20px; padding: 20px;">
Say the true data generating process is: $y_i = b_0 + b \cdot x_i + \epsilon_i$. Our estimated model is $y_i=b\cdot x_i + \epsilon_i$. Then we get the expectation of $b$:<br />
<br />
$$ E[b] = E[(X^TX)^{-1} X^T y]$$<br />
$$ = E[(X^TX)^{-1} X^T (b_0 + Xb + \epsilon)]$$<br />
$$ = E[(X^TX)^{-1} X^Tb_0 + \underbrace{(X^TX)^{-1} X^TX}_{=\mathbf{1}}b + (X^TX)^{-1} X^T \epsilon]$$<br />
If $E[\epsilon] = 0$ as assumed in linear regression:<br />
$$ = E[(X^TX)^{-1} X^Tb_0] + b + 0$$<br />
<br />
Obviously $b$ is biased if neither $(X^TX)^{-1} X^T$ nor $b_0$ are equal to 0. See that this applies even when $b_0$ is not significant different from 0!<br />
<br /></div>
<br />
<h3>
<b>When to use a no-intercept regression</b></h3>
<div>
Basically there is only one reason to perform a regression without using the intercept: Whenever your model is used to describe a process which is known to have a zero-intercept. Examples will be presented in the last article of this series.<br />
<br />
So stay tuned!<br />
<br /></div>
<h3>
<b>Literature</b></h3>
[1.]<b> $R^2$-problem</b> on <b><a href="http://stats.stackexchange.com/">CrossValidated</a></b>: <a href="http://stats.stackexchange.com/questions/26176/removal-of-statistically-significant-intercept-term-boosts-r2-in-linear-model/26205#26205">Link</a>.<br />
<div style="text-align: left;">
[2.] <b>William Greene:</b> <i>Econometrics </i>(<a href="http://people.stern.nyu.edu/wgreene/MathStat/GreeneChapter4.pdf">Link</a>; Biasedness of the OLS-estimator for omitted variables in Sections 4.3.2)</div>
<br />Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-45785504990836548452013-01-08T10:10:00.000+01:002013-01-09T11:11:10.724+01:00A Discussion on Non-Hierarchical Regression Models<br />
There is a certain point in many trainings where we talk about the advice to respect "model hierarchy" in the context of multiple linear regression models. Model hierarchy means: If there is a term of higher order included in the model, all corresponding terms of lower order should be in the model, too.<br />
<b><br /></b>
<b>Example</b><br />
If you estimated a model $Y = b_0 + b_1 X_1 + b_2 X_2 + b_{12} X_1 \cdot X_2 + \epsilon$ and see that $b_{12}$ is significant different from 0 the main effects $X_1$ and $X_2$ should stay in the model in any case. No matter if they are significant themselves or not.<br />
<br />
In the following series of articles I want to give an explanation why we give this recommendation.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhVcGE1eJJa_bA97VuTgn6yUcdQCGaMw9Hn2AbjQy1vOG5T5K1g3UV-iKzeSdhY8ciVq1fwhd9WZBkXjc2QJr-YnXzf-gL8F3Asa9jd91o0FEGmDlle09aCruQ9bgpgVlUxJ2Ee4IwYJgI/s1600/Reihenfolge.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="119" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhVcGE1eJJa_bA97VuTgn6yUcdQCGaMw9Hn2AbjQy1vOG5T5K1g3UV-iKzeSdhY8ciVq1fwhd9WZBkXjc2QJr-YnXzf-gL8F3Asa9jd91o0FEGmDlle09aCruQ9bgpgVlUxJ2Ee4IwYJgI/s400/Reihenfolge.png" width="400" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
<br />
In the <b><a href="http://statistiksoftware.blogspot.de/2013/01/why-we-need-intercept.html">first article</a></b> I will present the problems that occur if the intercept term drops out of the model. Of course this is the most simple kind of violating the model hierarchy. I give a motivation why we should prefer:<br />
<br />
$$ y = b + b_1 x_1 + \epsilon $$<br />
<br />
in nearly all cases over:<br />
<br />
$$ y = b_1 x_1 + \epsilon $$<br />
<br />
The <b>second article</b> will explain the difficulties that occur whenever you estimate a non-hierarchical model by removing terms of lower order, while keeping the corresponding terms of higher order. This includes not only 2 (and more)-factor-interactions but polynomial terms as well. So our general advice applies to models like<br />
<br />
$$ y = b + b_1 x_1 + b_{11} x_1^2 + \epsilon$$<br />
<br />
even though.<br />
<br />
Of course there are exceptions from these general rules. The <b>last article</b> will give some examples, where one might use non-hierarchical models.<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
</div>
Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-48498682339695908182012-09-05T10:09:00.001+02:002013-01-08T11:09:37.396+01:00Die R-Schnittstelle von JMP angewendetIm folgenden Beitrag erklärt unser Gastautor <b>Mirko Löhmann</b>, die "Auswertung von Objekten in der Ebene mit Räumlicher Statistik". Besonders spannend: Dabei werden die Vorzüge von JMP und R über die R-Schnittstelle von JMP benutzt.<br />
<div>
<br /></div>
<div>
<a href="http://www.statcon.de/blog/JMP_R_spatstat.pdf">www.statcon.de/blog/JMP_R_spatstat.pdf</a></div>
<div>
<br /></div>
<div>
Vielen Dank für diesen Beitrag!<br />
<br /></div>
Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-69215081971243413202012-05-11T11:41:00.000+02:002012-05-11T11:42:11.563+02:00Control 2012 - Tag 3<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi6STUCDnC__JovEWnMr0u70xMPtyJAZiimIAXKbjs7cci0GTtsn_4y2mgkxZeILieKSoRq-iaQtzGM-fnFmK7LdD6q7DOD7cheVfwLtaepjAGgjnfQnMapZi8r9-1QjboFA3i1948CWKQ/s1600/Tag+3+-+Messegeba%CC%88ude.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="312" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi6STUCDnC__JovEWnMr0u70xMPtyJAZiimIAXKbjs7cci0GTtsn_4y2mgkxZeILieKSoRq-iaQtzGM-fnFmK7LdD6q7DOD7cheVfwLtaepjAGgjnfQnMapZi8r9-1QjboFA3i1948CWKQ/s640/Tag+3+-+Messegeba%CC%88ude.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Das Messegebäude "von Hinten"</td></tr>
</tbody></table>
8:40 - Ich bin alleine am Stand und vervollständige die Notizen für den Blogeintrag "Tag 2". Von den Kollegen ist bisher niemand zu sehen. Wahrscheinlich bereiten sie gerade die Motivations-Ansprache vor. Während ich schreibe und und warte nehme ich mir vor mehr Notizen zu machen. Ich werde das Gefühl nicht los, noch etwas von gestern vergessen zu haben.<br />
<br />
8:55 - Ich bin nicht mehr alleine und auf Anfrage hält Doug tatsächlich eine, nicht ganz ernst gemeinte, Morivationsrede.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhK-feZQoQ9ft3QiMoeZLVtNUV9K_XCdH1HUem3iJF7mOoAktsAbs_ua2tFdr8bPiS5l9ZVtRZRs1Gc-1cR_MHALhc0zMej45BLxdk-4KQA9Q_QYP9uinz0z8bgZg9nq1OqgC73wGB0f3I/s1600/Tag+3+-+Doug+and+Mike.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhK-feZQoQ9ft3QiMoeZLVtNUV9K_XCdH1HUem3iJF7mOoAktsAbs_ua2tFdr8bPiS5l9ZVtRZRs1Gc-1cR_MHALhc0zMej45BLxdk-4KQA9Q_QYP9uinz0z8bgZg9nq1OqgC73wGB0f3I/s640/Tag+3+-+Doug+and+Mike.JPG" width="374" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Unsere Partner aus Amerika: Mike und Doug</td></tr>
</tbody></table>
9:15 - Das Highlight des Tages bahnt sich an: Nachdem Bertram das Tablet mit ins Hotel genommen und dort zahlreiche Wifi-Analyser-Apps installiert hat lebt die Hoffnung auf eine stabile Internetverbindung wieder auf.<br />
<br />
10:37 - Die App funktioniert. Da der Rest des Tages relativ ruhig verlief, ich es aber gleichzeitig nicht geschafft habe die letzte bisher unerschlossene Halle (5) zu besuchen, möchte ich kurz diese ominöse Android-App beschreiben:<br />
<br />
<b><u>Idee:</u></b><br />
Wir wollen Daten mit den Sensoren eines Android-Tablets erheben und mithilfe von InfinityQS analysieren. Das beinhaltet im Wesentlichen:<br />
<ol>
<li>Eine App, die die Daten erhebt, darstellt und an InfinityQS schickt.</li>
<li>Eine Schnittstellt zwischen der Android-App und InifinityQS zum Übertragen der Daten.</li>
</ol>
<u><b>Die App:</b></u><br />
Typische Android-Devices haben einen Lage-, Licht- und einen Abstandssensor. Ausserdem können natürlich Audio-Daten vom Mikrophon aufgenommen werden. Wir haben uns entschlossen den Lichtsensor sowie das Mikrophon zu verwenden. Dabei werden konstant Audio-Daten aufgenommen. Über eine <a href="http://de.wikipedia.org/wiki/Schnelle_Fourier-Transformation">Fast-Fourier-Transformation</a> werden diese Daten in Frequenzen und die zugehörigen Amplituden umgerechnet.Ausgewählte Frequenzen werden im Tablet grafisch dargestellt.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPNd2QG4HN5az2-wKgabP1FCNr6yw8Ewdx5CWuI-kW4ESvLP8TJ6qksJap7Cep4-gI9hb4L9KcG9qZpHhKVHLPYvdtZVJ0hyphenhyphenhwYCInnI_h5BiNnBcgP-W9xaVfizLlI_ItyCPKQtf1Q5k/s1600/image3063.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="480" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPNd2QG4HN5az2-wKgabP1FCNr6yw8Ewdx5CWuI-kW4ESvLP8TJ6qksJap7Cep4-gI9hb4L9KcG9qZpHhKVHLPYvdtZVJ0hyphenhyphenhwYCInnI_h5BiNnBcgP-W9xaVfizLlI_ItyCPKQtf1Q5k/s640/image3063.png" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Die Messeapp mit der Auswertung der Daten in InfinityQS</td></tr>
</tbody></table>
<u><b>Die Schnittstelle und InfinityQS:</b></u><br />
Auf dem Rechner werden die Daten von einem Java-Programm über eine TCP-IP-Verbindung empfangen und in eine Textdatei geschrieben. InfinityQS liest die Werte aus der Textdatei aus und analysiert sie anschließend. Im Bild sieht man ausgewählte Regelkarten für verschiedene Frequenze.<br />
<br />
Man kann an den Grafiken und Daten relative gut ablesen, ob gerade jemand vor dem Tablet steht und spricht oder nicht. Man kann von der gemessenen Lautstärke auch gut auf den <i>Traffic </i>vor dem Stand schließen.<br />
<br />
Am letzten Tag steht jetzt noch ein Ausflug in Halle 5 an. Dann werde ich hoffentlich noch viele interessante Bilder von Ausstellungsstücken präsentieren können!Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-86946834037061567812012-05-10T09:21:00.001+02:002012-05-10T11:23:11.252+02:00Control 2012 - Tag 2<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEittkYfrUrA4TjDBJ7E7nISQa8inau7o5MDxO-KLZ5TccT8Ec1HLVtvLuTMA6jg-7kZXNO9lBipNz1hqdw8vhmESz3Vmg3O0Kd132Uc4qf0UHw0GtgFZxTQZGEuMFgcv9prusr8KlLhRNg/s1600/Tag+2-+Messegeb%C3%A4ude.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="250" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEittkYfrUrA4TjDBJ7E7nISQa8inau7o5MDxO-KLZ5TccT8Ec1HLVtvLuTMA6jg-7kZXNO9lBipNz1hqdw8vhmESz3Vmg3O0Kd132Uc4qf0UHw0GtgFZxTQZGEuMFgcv9prusr8KlLhRNg/s640/Tag+2-+Messegeb%C3%A4ude.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Messegebäude andere Ansicht</td></tr>
</tbody></table>
<br />
08:30 - Da ich meine Eintrittskarte behalten habe läuft heute alles wie geschmiert. Um 9:00 Uhr stelle ich fest, dass diese Erwartungshaltung etwas voreilig war. Nachdem ich unsere Messe-App gestern Abend Zuhause noch einmal ausprobiert habe und alles stabil lief war ich optimistisch, dass wir das heute hinbekommen. Ausserdem haben Boris und Bertram ein neues Wunderwerk der Technik (W-Lan Router) mitgebracht. <br />
<br />
08:45 Mit einem hauch von Neid bemerken wir, das am Nachbarstand eine Motivationsrede für einen Erfolgreichen Arbeitstag gehalten wird. Das sollten wir auch einführen. Auch unser amerikanischer Kollege Doug ist begeistert. Ich bin gespannt, wer den Part morgen übernimmt.<br />
<br />
10:00 - Die Zeit vergeht wie im Fluge, da heute etwas mehr los zu sein scheint als gestern. <br />
<br />
11:30 - Die erste Packung mit Keksen wird ausgepackt. Wirklich lecker.<br />
<br />
14:30 - Wir stellen fest, dass der W-Lan Router sich positiv auf alle Geräte ausgewirkt hat. Die Verbindungen sind stabiler und/oder schneller. Leider gilt das mit einer Einschränkung: Galaxy Tab. Ich verbeuge mich vor der Technik und gebe auf. Zumindest produziert die App hübsche Bildchen auf dem Display (Foto folgt).<br />
<br />
14:30 - Ich mache mich auf um einige Fotos von Ausstellungsstücken nachzuholen, die mir gestern aufgefallen waren. Resultat:<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhJ6KRtKBZ9lASghViOFdUz0BmAcbernkMRicGeZR15B9RRLBFXRwlw1ZUdYsF0H3GbQRKtMOqq2-BieTBEeBOvI68KhlLks-JJ0RYYhey4pLD_S0NMYweREj2M5Vs-PECIfZK7M6bIN0Q/s1600/Tag+2+-+Nachtrag+1.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhJ6KRtKBZ9lASghViOFdUz0BmAcbernkMRicGeZR15B9RRLBFXRwlw1ZUdYsF0H3GbQRKtMOqq2-BieTBEeBOvI68KhlLks-JJ0RYYhey4pLD_S0NMYweREj2M5Vs-PECIfZK7M6bIN0Q/s640/Tag+2+-+Nachtrag+1.JPG" width="480" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Maschine zum Sortieren von Metallkugeln</td></tr>
</tbody></table>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhf384Dxd4er_I9SXBz40rZJbfWxc4QYJGhyphenhyphen0MMfUNfzTApArddNRYnDZM6gZVxKgMPUOD3dBj8vlPodXtNvjYqnJMsxHb3_Q0kL8kIksnMy-m3AU8-i76i1cmEWlUJJVajzjF9D9qM6cc/s1600/Tag+2+-+Nachtrag+2.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhf384Dxd4er_I9SXBz40rZJbfWxc4QYJGhyphenhyphen0MMfUNfzTApArddNRYnDZM6gZVxKgMPUOD3dBj8vlPodXtNvjYqnJMsxHb3_Q0kL8kIksnMy-m3AU8-i76i1cmEWlUJJVajzjF9D9qM6cc/s640/Tag+2+-+Nachtrag+2.JPG" width="484" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Projektor-R2-D2</td></tr>
</tbody></table>
15:30 - Der Rundgang durch Halle 7 mit Claudia und Doug zeigt, dass es dort nicht so spannend ist, wie bei uns in Halle 3. Es ist recht wenig los und die Aussteller haben sich zum großen Teil nicht so spannende Highlights überlegt. Die große Ausnahme der Regel folgt auf dem Fuße:<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjnRwMEX7ZS23cKt8iCMojO3MVdxLmPNfCWZfME_G4LgN62D3sqToVC0kZlMeOsF9gx-hChUVLKWxevev0RbloA91B-bFYs-h5C-pG1fHQ42MN18ivSXPKnVGg9vGG2G3vjmCVDadoZWmA/s1600/Tag+2+-+Highlight.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="480" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjnRwMEX7ZS23cKt8iCMojO3MVdxLmPNfCWZfME_G4LgN62D3sqToVC0kZlMeOsF9gx-hChUVLKWxevev0RbloA91B-bFYs-h5C-pG1fHQ42MN18ivSXPKnVGg9vGG2G3vjmCVDadoZWmA/s640/Tag+2+-+Highlight.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Highlight von Halle 7</td></tr>
</tbody></table>
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEireaLGkFPEvb8HUTHZDIT1rfkN2J31ijKfJf_Tx6voQC1pG4xQJC4LeMEx_uFFfcudj8IQDoQgpc0K8TOHydlLMcFXmCJJH8NCkoIR3lyD_Xb45UrXvzRaNtkydEcpE4o05frrfiLl3-g/s1600/Tag+2+-+Auto+2.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="368" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEireaLGkFPEvb8HUTHZDIT1rfkN2J31ijKfJf_Tx6voQC1pG4xQJC4LeMEx_uFFfcudj8IQDoQgpc0K8TOHydlLMcFXmCJJH8NCkoIR3lyD_Xb45UrXvzRaNtkydEcpE4o05frrfiLl3-g/s640/Tag+2+-+Auto+2.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Schokoladenauto? Ich durfte es nicht probieren!</td></tr>
</tbody></table>
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiW6OEDvq9rPNc75Fq5VE-rN37-qaavxxcll7047j8WJW51wkw0tT6euHWKsSWEcHaE1DMP-avIU4WSuLJ4qUPJ1Xph7rlZkruq9TN03umAE2rG5QSpCtoqmhCrysMRy4g-a9SR1Hltlqs/s1600/Tag+2+-+Auto.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="480" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiW6OEDvq9rPNc75Fq5VE-rN37-qaavxxcll7047j8WJW51wkw0tT6euHWKsSWEcHaE1DMP-avIU4WSuLJ4qUPJ1Xph7rlZkruq9TN03umAE2rG5QSpCtoqmhCrysMRy4g-a9SR1Hltlqs/s640/Tag+2+-+Auto.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Autos sind sehr beliebte Ausstellungsstücke - liegt das vielleicht an der Männerquote?</td></tr>
</tbody></table>
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEispfIW_C4HoRjF3_ChfgYjj7LbQ7TVur8m7zQI-N69G11XPIqUkjUgbiDahUoOuYJX_uXCHMIKM2vjz8axmZz7HKU9WxkgONSBmioATm5WGDravEPoQXBfGu60dOw96uOhd3E2NncYl2A/s1600/Tag+2+-+Erholungsoase.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="640" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEispfIW_C4HoRjF3_ChfgYjj7LbQ7TVur8m7zQI-N69G11XPIqUkjUgbiDahUoOuYJX_uXCHMIKM2vjz8axmZz7HKU9WxkgONSBmioATm5WGDravEPoQXBfGu60dOw96uOhd3E2NncYl2A/s640/Tag+2+-+Erholungsoase.JPG" width="480" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Die Erholungsoase. Aber Vorsicht: Keine Selbstbedienung</td></tr>
</tbody></table>
<br />
17:00 - Fazit: Heute war sicher mehr los. Aber das ganze kann sich gerne noch einmal ein bisschen Steigern in den nächsten Tagen.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjTQ6huvCM3IFL5bpffuyQJMmacrousgY8kJfATArJTlFCUjzqqQ2V_7OwXEL9M-XG8OC25HKLnTl1Th1b_0N3_EyuCeV0xKWwUC_TaP3LptexsCGy-6j_p5cm8U1vQyL93XcdvvgR9MGU/s1600/Tag+2+-+Flieger.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="456" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjTQ6huvCM3IFL5bpffuyQJMmacrousgY8kJfATArJTlFCUjzqqQ2V_7OwXEL9M-XG8OC25HKLnTl1Th1b_0N3_EyuCeV0xKWwUC_TaP3LptexsCGy-6j_p5cm8U1vQyL93XcdvvgR9MGU/s640/Tag+2+-+Flieger.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Hatte ich vergessen zu erwähnen, dass die Messe direkt am Flughafen liegt?</td></tr>
</tbody></table>
<br />
<br />Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-52091142202611253662012-05-08T20:36:00.001+02:002012-05-10T11:22:57.643+02:00Control 2012 - Erster Tag<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVqG9cB6xDm0VuTRSRVDX1U6ji7baYUBNNvcu3bd9CD2w8b5bh252A7CpFtC6nPAGl1_TVkS3eRyGWcvGBoDB5jJ9j3JbjerN9gmGFLoxkJErpKsz241762lXSX_aXLzyhKKE3Y8I5Lb0/s1600/Die+Messe.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="283" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVqG9cB6xDm0VuTRSRVDX1U6ji7baYUBNNvcu3bd9CD2w8b5bh252A7CpFtC6nPAGl1_TVkS3eRyGWcvGBoDB5jJ9j3JbjerN9gmGFLoxkJErpKsz241762lXSX_aXLzyhKKE3Y8I5Lb0/s640/Die+Messe.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Stuttgarter Messegebäude samt Parkplatz</td></tr>
</tbody></table>
<br />
8:30 - Pünktlich stehe ich vor dem Messegelände und erfahre, dass der Rest der Belegschaft noch unterwegs ist. Da ich als einziger nicht mit einer Eintrittskarte ausgestattet bin vertreibe ich mir die Zeit vor dem Messegelände und genieße das schöne Wetter.<br />
<br />
8:45 - Claudia versorgt mich mit der lang ersehnten Eintrittskarte und so kann es losgehen. Der Stand befindet sich unweit von der Stelle an der wir vor einem Jahr residierten. Das ist aber schon die einzige Gemeinsamkeit. Jetzt sind die Monitore professionell an den Wänden angebracht, die Plakate an den Standwänden überzeugen mit schlichter Eleganz und berücksichtigt man die einheitliche Oberbekleidung des Personals kann man das als einen Quantensprung in Sachen professioneller Erscheinung zusammenfassen.<br />
<br />
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj3XNTx8cPM1ItLVU1NXgHjuYpM8i6i4VSBHHMsDLmCwMetqL5t47NJSC6gC2HuHE9YaAwgnK2_kQspbh4rnv8eI2f5WPbLhvHXG4WJP1mcJ74Whx-wJkmU57fh8JAaYlBDdDukP8wyEEA/s1600/Der+Stand.JPG" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="480" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj3XNTx8cPM1ItLVU1NXgHjuYpM8i6i4VSBHHMsDLmCwMetqL5t47NJSC6gC2HuHE9YaAwgnK2_kQspbh4rnv8eI2f5WPbLhvHXG4WJP1mcJ74Whx-wJkmU57fh8JAaYlBDdDukP8wyEEA/s640/Der+Stand.JPG" width="640" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Messestand mit dem Statcon-Team (vlnr: Boris Kulig, Claudia Walber, Sebastian Hoffmeister, Bertram Schäfer)</td></tr>
</tbody></table>
9:20 - Der Eindruck übermäßiger Professionalität wird getrübt durch einen Mangel an funktionstüchtigem W-Lan. Leider ist eine Internetverbindung essentiell zum Betreiben der vorbereiteten Android-App. Ein Thema, dass mich den Rest des (nicht nur Messe-) Tages beschäftigen wird.<br />
<br />
11:00 - Ein Rundgang durch Halle 3 liefert zahlreiche neue Ideen, was man nächstes Jahr noch machen könnte. Angefangen von einer Maschine zum sortieren von Stahl-Kugeln gemäß ihres Durchmessers bis hin zur Liveproduktion von Schiedsrichterpfeifen gibt es allerlei praktische oder zumindest faszinierende Dinge zu bestaunen.<br />
<br />
12:00 - Nach einer Sonderzahlung an eine der beiden Firmen, die für die Internetverbindung vor Ort zuständig sind, läuft auch das W-Lan … auf einigen Geräten. Leider gilt das nicht für das Galaxy Tab, für das besagte Android-App entwickelt wurde.<br />
<br />
13:00 - Ein anderes Consulting und Trainings-Unternehmen beeindruckt durch einen R2-D2 in (annähernder) Lebensgröße. Zwar ist der Funktionsumfang eingeschränkt - statt Raumschiffe reparieren dient er nur als Projektor - aber er kann sich bewegen und lockt so sicher den einen oder anderen Kunden zum Stand.<br />
<br />
16:00 - Der Messetag geht relativ beschaulich zu Ende. Der Besucherandrang lässt sich wahrscheinlich an zwei Händen abzählen, aber das war am ersten Tag auch schon letztes Jahr der Fall.<br />
<br />
Ziel für Morgen:<br />
- viele Fotos von den spannenden Maschinen (präziser: Messmitteln) machenUnknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-28788110645283786352012-01-10T17:52:00.000+01:002012-01-16T17:46:33.572+01:00Interaktive Graphiken online PräsentierenWieder einmal gibt es eine neue Ausgabe des R Journal und wieder finden sich dort spannende Themen. Besonders aufgefallen ist mir dieses mal ein Artikel von Markus Gesmann und Diego de Castillo: <a href="http://journal.r-project.org/archive/2011-2/RJournal_2011-2_Gesmann+de~Castillo.pdf">Using the Google Visualisation API with R</a>!<br />
<br />
Mit dem <a href="http://cran.r-project.org/web/packages/googleVis/index.html">googleVis</a> Paket können interaktive Graphiken erstellt und direkt im Browser angezeigt werden.<br />
<br />
Als kleiner Vorgeschmack:<br />
<br />
<div style="text-align: center;">
<b>Die Wechselkurse von EUR, GBP, YEN und <a href="http://de.wikipedia.org/wiki/Renminbi">Renminbi</a> (Chinesische Währung) - jeweils zum USD.</b><title>AnnotatedTimeLineIDf456e3929f</title><style type="text/css">
body {
color: #444444;
font-family: Arial,Helvetica,sans-serif;
font-size: 75%;
}
a {
color: #4D87C7;
text-decoration: none;
}
</style><script src="http://www.google.com/jsapi" type="text/javascript">
</script><script type="text/javascript">
// jsData
function gvisDataAnnotatedTimeLineIDf456e3929f ()
{
var data = new google.visualization.DataTable();
var datajson =
[
[
new Date(2012,0,8),
1.2893,
null,
null,
1.5527,
null,
null,
0.013,
null,
null,
0.1576,
null,
null
],
[
new Date(2012,0,1),
1.3004,
null,
null,
1.5553,
null,
null,
0.0129,
null,
null,
0.157,
null,
null
],
[
new Date(2011,11,25),
1.3049,
null,
null,
1.5605,
null,
null,
0.0128,
null,
null,
0.1568,
null,
null
],
[
new Date(2011,11,18),
1.313,
null,
null,
1.5554,
null,
null,
0.0128,
null,
null,
0.1564,
"Nord-Korea",
"Kim Jong-Il stirbt."
],
[
new Date(2011,11,11),
1.3389,
null,
null,
1.5639,
null,
null,
0.0129,
null,
null,
0.1573,
null,
null
],
[
new Date(2011,11,4),
1.3367,
null,
null,
1.5585,
null,
null,
0.0128,
null,
null,
0.1568,
null,
null
],
[
new Date(2011,10,27),
1.3404,
null,
null,
1.5594,
null,
null,
0.013,
null,
null,
0.1567,
null,
null
],
[
new Date(2011,10,20),
1.3576,
null,
null,
1.5863,
null,
null,
0.013,
null,
null,
0.1573,
null,
null
],
[
new Date(2011,10,13),
1.3716,
null,
null,
1.6013,
null,
null,
0.0129,
null,
null,
0.1574,
null,
null
],
[
new Date(2011,10,6),
1.3854,
null,
null,
1.6023,
null,
null,
0.0129,
null,
null,
0.1571,
null,
null
],
[
new Date(2011,9,30),
1.3991,
null,
null,
1.6019,
null,
null,
0.0132,
null,
null,
0.1568,
null,
null
],
[
new Date(2011,9,23),
1.3808,
null,
null,
1.5808,
null,
null,
0.013,
null,
null,
0.1563,
null,
null
],
[
new Date(2011,9,16),
1.3671,
null,
null,
1.5684,
null,
null,
0.013,
null,
null,
0.1566,
null,
null
],
[
new Date(2011,9,9),
1.3341,
null,
null,
1.5491,
null,
null,
0.013,
null,
null,
0.1564,
null,
null
],
[
new Date(2011,9,2),
1.3514,
null,
null,
1.5568,
null,
null,
0.013,
null,
null,
0.156,
null,
null
],
[
new Date(2011,8,25),
1.3615,
null,
null,
1.5593,
null,
null,
0.0131,
null,
null,
0.1561,
null,
null
],
[
new Date(2011,8,18),
1.3715,
null,
null,
1.5811,
null,
null,
0.013,
null,
null,
0.1563,
null,
null
],
[
new Date(2011,8,11),
1.399,
null,
null,
1.6023,
null,
null,
0.0129,
null,
null,
0.1563,
null,
null
],
[
new Date(2011,8,4),
1.4378,
null,
null,
1.6287,
null,
null,
0.013,
null,
null,
0.1565,
null,
null
],
[
new Date(2011,7,28),
1.4419,
null,
null,
1.6415,
null,
null,
0.013,
null,
null,
0.1564,
null,
null
],
[
new Date(2011,7,21),
1.4361,
null,
null,
1.6419,
null,
null,
0.013,
null,
null,
0.1566,
null,
null
],
[
new Date(2011,7,14),
1.4254,
null,
null,
1.6287,
null,
null,
0.0129,
null,
null,
0.1558,
null,
null
],
[
new Date(2011,7,7),
1.4271,
null,
null,
1.6352,
null,
null,
0.0129,
null,
null,
0.1553,
null,
null
],
[
new Date(2011,6,31),
1.4386,
null,
null,
1.635,
null,
null,
0.0128,
null,
null,
0.1551,
null,
null
],
[
new Date(2011,6,24),
1.4225,
null,
null,
1.6178,
null,
null,
0.0127,
null,
null,
0.1549,
null,
null
],
[
new Date(2011,6,17),
1.413,
null,
null,
1.604,
null,
null,
0.0126,
null,
null,
0.1548,
null,
null
],
[
new Date(2011,6,10),
1.4399,
null,
null,
1.6039,
null,
null,
0.0124,
null,
null,
0.1546,
null,
null
],
[
new Date(2011,6,3),
1.437,
null,
null,
1.6011,
null,
null,
0.0124,
null,
null,
0.1545,
null,
null
],
[
new Date(2011,5,26),
1.4282,
null,
null,
1.61,
null,
null,
0.0125,
null,
null,
0.1544,
null,
null
],
[
new Date(2011,5,19),
1.4308,
null,
null,
1.624,
null,
null,
0.0124,
null,
null,
0.1542,
null,
null
],
[
new Date(2011,5,12),
1.4557,
null,
null,
1.6363,
null,
null,
0.0125,
null,
null,
0.1541,
null,
null
],
[
new Date(2011,5,5),
1.442,
null,
null,
1.6431,
null,
null,
0.0124,
null,
null,
0.1541,
null,
null
],
[
new Date(2011,4,29),
1.4147,
null,
null,
1.6285,
null,
null,
0.0123,
null,
null,
0.1538,
null,
null
],
[
new Date(2011,4,22),
1.4193,
null,
null,
1.6207,
null,
null,
0.0123,
null,
null,
0.1536,
null,
null
],
[
new Date(2011,4,15),
1.4273,
null,
null,
1.632,
null,
null,
0.0124,
null,
null,
0.1538,
null,
null
],
[
new Date(2011,4,8),
1.4694,
null,
null,
1.6525,
null,
null,
0.0124,
null,
null,
0.1538,
null,
null
],
[
new Date(2011,4,1),
1.4696,
null,
null,
1.6578,
null,
null,
0.0122,
null,
null,
0.1536,
null,
null
],
[
new Date(2011,3,24),
1.445,
null,
null,
1.6393,
null,
null,
0.0121,
null,
null,
0.1532,
null,
null
],
[
new Date(2011,3,17),
1.4454,
null,
null,
1.6323,
null,
null,
0.0119,
null,
null,
0.1528,
null,
null
],
[
new Date(2011,3,10),
1.43,
null,
null,
1.6257,
null,
null,
0.0118,
null,
null,
0.1526,
null,
null
],
[
new Date(2011,3,3),
1.4128,
null,
null,
1.6043,
null,
null,
0.0121,
null,
null,
0.1523,
null,
null
],
[
new Date(2011,2,27),
1.4155,
null,
null,
1.6208,
null,
null,
0.0123,
null,
null,
0.1522,
null,
null
],
[
new Date(2011,2,20),
1.3999,
null,
null,
1.6114,
null,
null,
0.0124,
null,
null,
0.1519,
null,
null
],
[
new Date(2011,2,13),
1.3911,
null,
null,
1.6163,
null,
null,
0.0121,
"Japan",
"Beginn der Fukushima-Katastrophe",
0.1521,
null,
null
],
[
new Date(2011,2,6),
1.3855,
null,
null,
1.6238,
null,
null,
0.0122,
null,
null,
0.152,
null,
null
],
[
new Date(2011,1,27),
1.3718,
null,
null,
1.6179,
null,
null,
0.0121,
null,
null,
0.1519,
null,
null
],
[
new Date(2011,1,20),
1.3565,
null,
null,
1.6112,
null,
null,
0.012,
null,
null,
0.1516,
null,
null
],
[
new Date(2011,1,13),
1.3599,
null,
null,
1.6075,
null,
null,
0.0121,
null,
null,
0.1518,
null,
null
],
[
new Date(2011,1,6),
1.3678,
null,
null,
1.6058,
null,
null,
0.0122,
null,
null,
0.1519,
null,
null
],
[
new Date(2011,0,30),
1.365,
null,
null,
1.5908,
null,
null,
0.0121,
null,
null,
0.1517,
null,
null
],
[
new Date(2011,0,23),
1.3443,
null,
null,
1.5937,
null,
null,
0.0121,
null,
null,
0.1517,
null,
null
],
[
new Date(2011,0,16),
1.3102,
null,
null,
1.5686,
null,
null,
0.012,
null,
null,
0.1511,
null,
null
],
[
new Date(2011,0,9),
1.3182,
null,
null,
1.5533,
null,
null,
0.0121,
null,
null,
0.1511,
null,
null
],
[
new Date(2011,0,2),
1.3222,
null,
null,
1.5467,
null,
null,
0.0122,
null,
null,
0.1511,
null,
null
],
[
new Date(2010,11,26),
1.3133,
null,
null,
1.5468,
null,
null,
0.012,
null,
null,
0.1501,
null,
null
],
[
new Date(2010,11,19),
1.3268,
null,
null,
1.5688,
null,
null,
0.0119,
null,
null,
0.15,
null,
null
],
[
new Date(2010,11,12),
1.3289,
null,
null,
1.5769,
null,
null,
0.012,
null,
null,
0.15,
null,
null
],
[
new Date(2010,11,5),
1.3198,
null,
null,
1.5618,
null,
null,
0.0119,
null,
null,
0.1499,
null,
null
],
[
new Date(2010,10,28),
1.3446,
null,
null,
1.5814,
null,
null,
0.012,
null,
null,
0.1501,
null,
null
],
[
new Date(2010,10,21),
1.3621,
null,
null,
1.6007,
null,
null,
0.012,
null,
null,
0.1504,
null,
null
],
[
new Date(2010,10,14),
1.3817,
null,
null,
1.6115,
null,
null,
0.0122,
null,
null,
0.1502,
null,
null
],
[
new Date(2010,10,7),
1.4034,
null,
null,
1.6111,
null,
null,
0.0124,
null,
null,
0.1497,
null,
null
],
[
new Date(2010,9,31),
1.3911,
null,
null,
1.5844,
null,
null,
0.0123,
null,
null,
0.1497,
null,
null
],
[
new Date(2010,9,24),
1.3923,
null,
null,
1.5794,
null,
null,
0.0123,
null,
null,
0.1502,
null,
null
],
[
new Date(2010,9,17),
1.3965,
null,
null,
1.5939,
null,
null,
0.0122,
null,
null,
0.15,
null,
null
],
[
new Date(2010,9,10),
1.3849,
null,
null,
1.588,
null,
null,
0.0121,
null,
null,
0.1494,
null,
null
],
[
new Date(2010,9,3),
1.3599,
null,
null,
1.5812,
null,
null,
0.0119,
null,
null,
0.1492,
null,
null
],
[
new Date(2010,8,26),
1.3268,
null,
null,
1.567,
null,
null,
0.0118,
null,
null,
0.1488,
null,
null
],
[
new Date(2010,8,19),
1.2944,
null,
null,
1.5526,
null,
null,
0.0118,
null,
null,
0.1481,
null,
null
],
[
new Date(2010,8,12),
1.2761,
null,
null,
1.5405,
null,
null,
0.0119,
null,
null,
0.1472,
null,
null
],
[
new Date(2010,8,5),
1.2782,
null,
null,
1.5445,
null,
null,
0.0118,
null,
null,
0.1468,
null,
null
],
[
new Date(2010,7,29),
1.2698,
null,
null,
1.55,
null,
null,
0.0118,
null,
null,
0.1469,
null,
null
],
[
new Date(2010,7,22),
1.2796,
null,
null,
1.5585,
null,
null,
0.0117,
null,
null,
0.147,
null,
null
],
[
new Date(2010,7,15),
1.3009,
null,
null,
1.5741,
null,
null,
0.0117,
null,
null,
0.1473,
null,
null
],
[
new Date(2010,7,8),
1.3177,
null,
null,
1.5868,
null,
null,
0.0116,
null,
null,
0.1474,
null,
null
],
[
new Date(2010,7,1),
1.2999,
null,
null,
1.5566,
null,
null,
0.0115,
null,
null,
0.1474,
null,
null
],
[
new Date(2010,6,25),
1.2894,
null,
null,
1.5296,
null,
null,
0.0115,
null,
null,
0.1473,
null,
null
],
[
new Date(2010,6,18),
1.2752,
null,
null,
1.5203,
null,
null,
0.0114,
null,
null,
0.1474,
null,
null
],
[
new Date(2010,6,11),
1.2607,
null,
null,
1.5141,
null,
null,
0.0114,
null,
null,
0.1474,
null,
null
],
[
new Date(2010,6,4),
1.2373,
null,
null,
1.5084,
null,
null,
0.0113,
null,
null,
0.1472,
null,
null
],
[
new Date(2010,5,27),
1.2334,
null,
null,
1.49,
null,
null,
0.0111,
null,
null,
0.1467,
null,
null
],
[
new Date(2010,5,20),
1.2286,
null,
null,
1.4747,
null,
null,
0.011,
null,
null,
0.1461,
null,
null
],
[
new Date(2010,5,13),
1.2017,
null,
null,
1.452,
null,
null,
0.0109,
null,
null,
0.1462,
null,
null
],
[
new Date(2010,5,6),
1.2186,
null,
null,
1.4551,
null,
null,
0.0109,
null,
null,
0.1462,
null,
null
],
[
new Date(2010,4,30),
1.2346,
null,
null,
1.4445,
null,
null,
0.0111,
null,
null,
0.1462,
null,
null
],
[
new Date(2010,4,23),
1.2402,
null,
null,
1.4422,
null,
null,
0.0109,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,4,16),
1.2636,
null,
null,
1.4758,
null,
null,
0.0108,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,4,9),
1.2957,
null,
null,
1.5046,
null,
null,
0.0107,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,4,2),
1.3289,
null,
null,
1.5316,
null,
null,
0.0106,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,3,25),
1.3408,
null,
null,
1.5358,
null,
null,
0.0107,
null,
null,
0.1465,
null,
null
],
[
new Date(2010,3,18),
1.356,
null,
null,
1.5407,
null,
null,
0.0108,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,3,11),
1.3426,
null,
null,
1.5263,
null,
null,
0.0107,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,3,4),
1.3483,
null,
null,
1.5111,
null,
null,
0.0107,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,2,28),
1.3431,
null,
null,
1.495,
null,
null,
0.0109,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,2,21),
1.367,
null,
null,
1.5156,
null,
null,
0.011,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,2,14),
1.3663,
null,
null,
1.5076,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,2,7),
1.3607,
null,
null,
1.5071,
null,
null,
0.0112,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,1,28),
1.358,
null,
null,
1.5369,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,1,21),
1.3618,
null,
null,
1.5616,
null,
null,
0.011,
null,
null,
0.1462,
null,
null
],
[
new Date(2010,1,14),
1.3687,
null,
null,
1.5647,
null,
null,
0.0112,
null,
null,
0.1462,
null,
null
],
[
new Date(2010,1,7),
1.3835,
null,
null,
1.5863,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,0,31),
1.4036,
null,
null,
1.6126,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,0,24),
1.4233,
null,
null,
1.6251,
null,
null,
0.011,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,0,17),
1.4459,
null,
null,
1.6191,
null,
null,
0.0109,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,0,10),
1.4363,
null,
null,
1.6041,
null,
null,
0.0108,
null,
null,
0.1463,
null,
null
],
[
new Date(2010,0,3),
1.4365,
null,
null,
1.6034,
null,
null,
0.0108,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,11,27),
1.4336,
null,
null,
1.6026,
null,
null,
0.011,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,11,20),
1.4497,
null,
null,
1.6229,
null,
null,
0.0112,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,11,13),
1.4751,
null,
null,
1.6332,
null,
null,
0.0112,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,11,6),
1.5013,
null,
null,
1.6544,
null,
null,
0.0114,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,10,29),
1.4968,
null,
null,
1.6545,
null,
null,
0.0114,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,10,22),
1.4906,
null,
null,
1.668,
null,
null,
0.0112,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,10,15),
1.4928,
null,
null,
1.6658,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,10,8),
1.479,
null,
null,
1.6491,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,10,1),
1.4844,
null,
null,
1.6391,
null,
null,
0.0109,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,9,25),
1.4965,
null,
null,
1.6424,
null,
null,
0.011,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,9,18),
1.4843,
null,
null,
1.6037,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,9,11),
1.469,
null,
null,
1.5941,
null,
null,
0.0112,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,9,4),
1.4606,
null,
null,
1.5937,
null,
null,
0.0111,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,8,27),
1.4714,
"Deutschland",
"Bundestagswahl: Union und FDP erhalten die Mehrheit.",
1.6187,
null,
null,
0.011,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,8,20),
1.4656,
null,
null,
1.6483,
null,
null,
0.011,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,8,13),
1.4468,
null,
null,
1.6523,
null,
null,
0.0109,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,8,6),
1.428,
null,
null,
1.6285,
null,
null,
0.0108,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,7,30),
1.4305,
null,
null,
1.6347,
null,
null,
0.0106,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,7,23),
1.4203,
null,
null,
1.6482,
null,
null,
0.0106,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,7,16),
1.4197,
null,
null,
1.6556,
null,
null,
0.0104,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,7,9),
1.4319,
null,
null,
1.6823,
null,
null,
0.0105,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,7,2),
1.4177,
null,
null,
1.65,
null,
null,
0.0105,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,6,26),
1.4188,
null,
null,
1.6436,
null,
null,
0.0106,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,6,19),
1.4029,
null,
null,
1.6307,
null,
null,
0.0107,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,6,12),
1.3945,
null,
null,
1.6212,
null,
null,
0.0106,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,6,5),
1.4044,
null,
null,
1.6451,
null,
null,
0.0104,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,5,28),
1.398,
null,
null,
1.644,
null,
null,
0.0104,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,5,21),
1.3916,
null,
null,
1.6393,
null,
null,
0.0103,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,5,14),
1.3997,
null,
null,
1.6257,
null,
null,
0.0102,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,5,7),
1.4145,
null,
null,
1.6258,
null,
null,
0.0104,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,4,31),
1.3995,
null,
null,
1.5989,
null,
null,
0.0105,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,4,24),
1.3713,
null,
null,
1.5559,
null,
null,
0.0105,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,4,17),
1.3596,
null,
null,
1.5192,
null,
null,
0.0104,
null,
null,
0.1464,
null,
null
],
[
new Date(2009,4,10),
1.3382,
null,
null,
1.5048,
null,
null,
0.0101,
null,
null,
0.1464,
null,
null
],
[
new Date(2009,4,3),
1.3206,
null,
null,
1.4738,
null,
null,
0.0102,
null,
null,
0.1463,
null,
null
],
[
new Date(2009,3,26),
1.3055,
null,
null,
1.4646,
null,
null,
0.0102,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,3,19),
1.3182,
null,
null,
1.4824,
null,
null,
0.01,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,3,12),
1.3299,
null,
null,
1.4726,
null,
null,
0.01,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,3,5),
1.332,
null,
null,
1.4479,
null,
null,
0.0101,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,2,29),
1.3515,
null,
null,
1.4506,
null,
null,
0.0103,
null,
null,
0.1462,
null,
null
],
[
new Date(2009,2,22),
1.324,
null,
null,
1.4202,
null,
null,
0.0103,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,2,15),
1.2763,
null,
null,
1.3919,
null,
null,
0.0102,
null,
null,
0.146,
null,
null
],
[
new Date(2009,2,8),
1.2609,
null,
null,
1.4143,
null,
null,
0.0102,
null,
null,
0.1459,
null,
null
],
[
new Date(2009,2,1),
1.2759,
null,
null,
1.4383,
null,
null,
0.0104,
null,
null,
0.146,
null,
null
],
[
new Date(2009,1,22),
1.2709,
null,
null,
1.4294,
null,
null,
0.0108,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,1,15),
1.2907,
null,
null,
1.4551,
null,
null,
0.011,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,1,8),
1.2857,
null,
null,
1.4499,
null,
null,
0.0111,
null,
null,
0.146,
null,
null
],
[
new Date(2009,1,1),
1.3025,
null,
null,
1.4136,
null,
null,
0.0112,
null,
null,
0.146,
null,
null
],
[
new Date(2009,0,25),
1.3045,
null,
null,
1.4113,
null,
null,
0.0111,
null,
null,
0.146,
null,
null
],
[
new Date(2009,0,18),
1.3286,
null,
null,
1.4785,
null,
null,
0.0111,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,0,11),
1.3638,
null,
null,
1.489,
null,
null,
0.0109,
null,
null,
0.1461,
null,
null
],
[
new Date(2009,0,4),
1.4022,
null,
null,
1.4556,
null,
null,
0.011,
null,
null,
0.1459,
null,
null
],
[
new Date(2008,11,28),
1.3988,
null,
null,
1.4767,
null,
null,
0.0111,
null,
null,
0.1457,
null,
null
],
[
new Date(2008,11,21),
1.3886,
null,
null,
1.5163,
null,
null,
0.0112,
null,
null,
0.1458,
null,
null
],
[
new Date(2008,11,14),
1.3036,
null,
null,
1.4832,
null,
null,
0.0109,
null,
null,
0.1455,
null,
null
],
[
new Date(2008,11,7),
1.2687,
null,
null,
1.4899,
null,
null,
0.0107,
null,
null,
0.1453,
null,
null
],
[
new Date(2008,10,30),
1.2797,
null,
null,
1.5224,
null,
null,
0.0105,
null,
null,
0.1462,
null,
null
],
[
new Date(2008,10,23),
1.2583,
null,
null,
1.4885,
null,
null,
0.0104,
null,
null,
0.1462,
null,
null
],
[
new Date(2008,10,16),
1.2661,
null,
null,
1.5241,
null,
null,
0.0103,
null,
null,
0.1462,
null,
null
],
[
new Date(2008,10,9),
1.2784,
null,
null,
1.5878,
null,
null,
0.0101,
null,
null,
0.146,
null,
null
],
[
new Date(2008,10,2),
1.2701,
null,
null,
1.5988,
null,
null,
0.0104,
null,
null,
0.1458,
null,
null
],
[
new Date(2008,9,26),
1.3022,
null,
null,
1.6566,
null,
null,
0.0102,
null,
null,
0.1459,
null,
null
],
[
new Date(2008,9,19),
1.3505,
null,
null,
1.7283,
null,
null,
0.0099,
null,
null,
0.1461,
null,
null
],
[
new Date(2008,9,12),
1.3592,
null,
null,
1.7353,
null,
null,
0.0098,
null,
null,
0.1462,
null,
null
],
[
new Date(2008,9,5),
1.4134,
null,
null,
1.7921,
null,
null,
0.0095,
null,
null,
0.1459,
null,
null
],
[
new Date(2008,8,28),
1.4623,
null,
null,
1.845,
null,
null,
0.0094,
null,
null,
0.1461,
null,
null
],
[
new Date(2008,8,21),
1.4294,
null,
null,
1.805,
null,
null,
0.0094,
null,
null,
0.1459,
null,
null
],
[
new Date(2008,8,14),
1.4147,
null,
null,
1.7675,
null,
null,
0.0093,
null,
null,
0.146,
null,
null
],
[
new Date(2008,8,7),
1.4473,
null,
null,
1.7847,
null,
null,
0.0093,
null,
null,
0.146,
null,
null
],
[
new Date(2008,7,31),
1.4722,
null,
null,
1.8381,
null,
null,
0.0091,
null,
null,
0.146,
null,
null
],
[
new Date(2008,7,24),
1.4754,
null,
null,
1.8634,
null,
null,
0.0091,
null,
null,
0.1457,
null,
null
],
[
new Date(2008,7,17),
1.4871,
null,
null,
1.8894,
null,
null,
0.0091,
null,
null,
0.1455,
null,
null
],
[
new Date(2008,7,10),
1.5385,
null,
null,
1.9503,
null,
null,
0.0092,
null,
null,
0.1457,
null,
null
],
[
new Date(2008,7,3),
1.5629,
null,
null,
1.9828,
null,
null,
0.0093,
null,
null,
0.1461,
null,
null
],
[
new Date(2008,6,27),
1.5776,
null,
null,
1.9944,
null,
null,
0.0093,
null,
null,
0.1463,
null,
null
],
[
new Date(2008,6,20),
1.5885,
null,
null,
1.9965,
null,
null,
0.0094,
null,
null,
0.1463,
null,
null
],
[
new Date(2008,6,13),
1.575,
null,
null,
1.9786,
null,
null,
0.0094,
null,
null,
0.1458,
null,
null
],
[
new Date(2008,6,6),
1.577,
null,
null,
1.9894,
null,
null,
0.0094,
null,
null,
0.1456,
null,
null
],
[
new Date(2008,5,29),
1.5646,
null,
null,
1.9775,
null,
null,
0.0093,
null,
null,
0.1454,
null,
null
],
[
new Date(2008,5,22),
1.5504,
null,
null,
1.9618,
null,
null,
0.0093,
null,
null,
0.145,
null,
null
],
[
new Date(2008,5,15),
1.5542,
null,
null,
1.9584,
null,
null,
0.0094,
null,
null,
0.1444,
null,
null
],
[
new Date(2008,5,8),
1.556,
null,
null,
1.9653,
null,
null,
0.0095,
null,
null,
0.144,
null,
null
],
[
new Date(2008,5,1),
1.5659,
null,
null,
1.9784,
null,
null,
0.0096,
null,
null,
0.1438,
null,
null
],
[
new Date(2008,4,25),
1.5671,
null,
null,
1.9673,
null,
null,
0.0096,
null,
null,
0.1435,
null,
null
],
[
new Date(2008,4,18),
1.5492,
null,
null,
1.9505,
null,
null,
0.0096,
null,
null,
0.1428,
null,
null
],
[
new Date(2008,4,11),
1.5449,
null,
null,
1.9626,
null,
null,
0.0096,
null,
null,
0.1428,
null,
null
],
[
new Date(2008,4,4),
1.5551,
null,
null,
1.9795,
null,
null,
0.0096,
null,
null,
0.1428,
null,
null
],
[
new Date(2008,3,27),
1.5801,
null,
null,
1.986,
null,
null,
0.0096,
null,
null,
0.1427,
null,
null
],
[
new Date(2008,3,20),
1.5835,
null,
null,
1.9784,
null,
null,
0.0098,
null,
null,
0.1428,
null,
null
],
[
new Date(2008,3,13),
1.5761,
null,
null,
1.9782,
null,
null,
0.0098,
null,
null,
0.1426,
null,
null
],
[
new Date(2008,3,6),
1.5712,
null,
null,
1.9886,
null,
null,
0.0099,
null,
null,
0.1423,
null,
null
],
[
new Date(2008,2,30),
1.5631,
null,
null,
1.9943,
null,
null,
0.01,
null,
null,
0.142,
null,
null
],
[
new Date(2008,2,23),
1.5611,
null,
null,
1.9979,
null,
null,
0.0101,
null,
null,
0.1412,
null,
null
],
[
new Date(2008,2,16),
1.5476,
null,
null,
2.0188,
null,
null,
0.0099,
null,
null,
0.1406,
null,
null
],
[
new Date(2008,2,9),
1.5261,
null,
null,
1.9951,
null,
null,
0.0097,
null,
null,
0.1405,
null,
null
],
[
new Date(2008,2,2),
1.5003,
null,
null,
1.9786,
null,
null,
0.0094,
null,
null,
0.14,
null,
null
],
[
new Date(2008,1,24),
1.4731,
null,
null,
1.9561,
null,
null,
0.0093,
null,
null,
0.1396,
null,
null
],
[
new Date(2008,1,17),
1.4581,
null,
null,
1.957,
null,
null,
0.0093,
null,
null,
0.1389,
null,
null
],
[
new Date(2008,1,10),
1.4653,
null,
null,
1.9587,
null,
null,
0.0094,
null,
null,
0.1389,
null,
null
],
[
new Date(2008,1,3),
1.4779,
null,
null,
1.9824,
null,
null,
0.0094,
null,
null,
0.1389,
null,
null
],
[
new Date(2008,0,27),
1.4612,
null,
null,
1.9614,
null,
null,
0.0094,
null,
null,
0.1382,
null,
null
],
[
new Date(2008,0,20),
1.4741,
null,
null,
1.9606,
null,
null,
0.0093,
null,
null,
0.1378,
null,
null
],
[
new Date(2008,0,13),
1.4731,
null,
null,
1.9654,
null,
null,
0.0092,
null,
null,
0.1374,
null,
null
],
[
new Date(2008,0,6),
1.4695,
null,
null,
1.9837,
null,
null,
0.009,
null,
null,
0.137,
null,
null
],
[
new Date(2007,11,30),
1.4498,
null,
null,
1.9858,
null,
null,
0.0088,
null,
null,
0.136,
null,
null
],
[
new Date(2007,11,23),
1.4388,
null,
null,
2.0029,
null,
null,
0.0088,
null,
null,
0.1355,
null,
null
],
[
new Date(2007,11,16),
1.4626,
null,
null,
2.0354,
null,
null,
0.0089,
null,
null,
0.1353,
null,
null
],
[
new Date(2007,11,9),
1.4656,
null,
null,
2.044,
null,
null,
0.009,
null,
null,
0.1349,
null,
null
],
[
new Date(2007,11,2),
1.4787,
null,
null,
2.0648,
null,
null,
0.0091,
null,
null,
0.1351,
null,
null
],
[
new Date(2007,10,25),
1.4771,
null,
null,
2.0592,
null,
null,
0.0091,
null,
null,
0.1347,
null,
null
],
[
new Date(2007,10,18),
1.4636,
null,
null,
2.0645,
null,
null,
0.009,
null,
null,
0.1345,
null,
null
],
[
new Date(2007,10,11),
1.4595,
null,
null,
2.0931,
null,
null,
0.0088,
null,
null,
0.1343,
null,
null
],
[
new Date(2007,10,4),
1.4439,
null,
null,
2.0707,
null,
null,
0.0087,
null,
null,
0.1337,
null,
null
],
[
new Date(2007,9,28),
1.4291,
null,
null,
2.0479,
null,
null,
0.0087,
null,
null,
0.1332,
null,
null
],
[
new Date(2007,9,21),
1.4223,
null,
null,
2.0407,
null,
null,
0.0086,
null,
null,
0.1329,
null,
null
],
[
new Date(2007,9,14),
1.4135,
null,
null,
2.037,
null,
null,
0.0085,
null,
null,
0.133,
null,
null
],
[
new Date(2007,9,7),
1.4175,
null,
null,
2.0407,
null,
null,
0.0086,
null,
null,
0.133,
null,
null
],
[
new Date(2007,8,30),
1.4144,
null,
null,
2.0246,
null,
null,
0.0087,
null,
null,
0.133,
null,
null
],
[
new Date(2007,8,23),
1.3966,
null,
null,
2.0069,
null,
null,
0.0087,
null,
null,
0.1329,
null,
null
],
[
new Date(2007,8,16),
1.3836,
null,
null,
2.024,
null,
null,
0.0088,
null,
null,
0.1328,
null,
null
],
[
new Date(2007,8,9),
1.3658,
null,
null,
2.0193,
null,
null,
0.0087,
null,
null,
0.1324,
null,
null
],
[
new Date(2007,8,2),
1.3644,
null,
null,
2.0128,
null,
null,
0.0086,
null,
null,
0.1322,
null,
null
],
[
new Date(2007,7,26),
1.3535,
null,
null,
1.9931,
null,
null,
0.0087,
null,
null,
0.1317,
null,
null
],
[
new Date(2007,7,19),
1.3539,
null,
null,
1.9976,
null,
null,
0.0086,
null,
null,
0.1316,
null,
null
],
[
new Date(2007,7,12),
1.375,
null,
null,
2.0296,
null,
null,
0.0084,
null,
null,
0.1319,
null,
null
],
[
new Date(2007,7,5),
1.3688,
null,
null,
2.0308,
null,
null,
0.0084,
null,
null,
0.132,
null,
null
],
[
new Date(2007,6,29),
1.3755,
null,
null,
2.049,
null,
null,
0.0083,
null,
null,
0.1321,
null,
null
],
[
new Date(2007,6,22),
1.3796,
null,
null,
2.0455,
null,
null,
0.0082,
null,
null,
0.132,
null,
null
],
[
new Date(2007,6,15),
1.3711,
null,
null,
2.0238,
null,
null,
0.0082,
null,
null,
0.1318,
null,
null
],
[
new Date(2007,6,8),
1.3596,
null,
null,
2.0127,
null,
null,
0.0081,
null,
null,
0.1314,
null,
null
],
[
new Date(2007,6,1),
1.3471,
null,
null,
2.0008,
null,
null,
0.0081,
null,
null,
0.1311,
null,
null
],
[
new Date(2007,5,24),
1.3413,
null,
null,
1.9882,
null,
null,
0.0081,
null,
null,
0.131,
null,
null
],
[
new Date(2007,5,17),
1.334,
null,
null,
1.9714,
null,
null,
0.0082,
null,
null,
0.1307,
null,
null
],
[
new Date(2007,5,10),
1.3453,
null,
null,
1.9832,
null,
null,
0.0082,
null,
null,
0.1306,
null,
null
],
[
new Date(2007,5,3),
1.3445,
null,
null,
1.9809,
null,
null,
0.0082,
null,
null,
0.1306,
null,
null
],
[
new Date(2007,4,27),
1.3461,
null,
null,
1.9789,
null,
null,
0.0082,
null,
null,
0.1304,
null,
null
],
[
new Date(2007,4,20),
1.3529,
null,
null,
1.9786,
null,
null,
0.0083,
null,
null,
0.1301,
null,
null
],
[
new Date(2007,4,13),
1.3549,
null,
null,
1.9887,
null,
null,
0.0083,
null,
null,
0.1298,
null,
null
],
[
new Date(2007,4,6),
1.3607,
null,
null,
1.9939,
null,
null,
0.0083,
null,
null,
0.1296,
null,
null
],
[
new Date(2007,3,29),
1.3611,
null,
null,
1.9994,
null,
null,
0.0084,
null,
null,
0.1293,
null,
null
],
[
new Date(2007,3,22),
1.3571,
null,
null,
1.9982,
null,
null,
0.0084,
null,
null,
0.1293,
null,
null
],
[
new Date(2007,3,15),
1.3439,
null,
null,
1.974,
null,
null,
0.0084,
null,
null,
0.1293,
null,
null
],
[
new Date(2007,3,8),
1.3369,
null,
null,
1.9713,
null,
null,
0.0084,
null,
null,
0.1292,
null,
null
],
[
new Date(2007,3,1),
1.3322,
null,
null,
1.9639,
null,
null,
0.0085,
null,
null,
0.1292,
null,
null
],
[
new Date(2007,2,25),
1.3313,
null,
null,
1.9556,
null,
null,
0.0085,
null,
null,
0.1292,
null,
null
],
[
new Date(2007,2,18),
1.3211,
null,
null,
1.9349,
null,
null,
0.0085,
null,
null,
0.129,
null,
null
],
[
new Date(2007,2,11),
1.3137,
null,
null,
1.9317,
null,
null,
0.0086,
null,
null,
0.129,
null,
null
],
[
new Date(2007,2,4),
1.3191,
null,
null,
1.9582,
null,
null,
0.0084,
null,
null,
0.129,
null,
null
],
[
new Date(2007,1,25),
1.3141,
null,
null,
1.9543,
null,
null,
0.0083,
null,
null,
0.1289,
null,
null
],
[
new Date(2007,1,18),
1.3065,
null,
null,
1.9509,
null,
null,
0.0083,
null,
null,
0.1288,
null,
null
],
[
new Date(2007,1,11),
1.298,
null,
null,
1.9614,
null,
null,
0.0083,
null,
null,
0.1288,
null,
null
],
[
new Date(2007,1,4),
1.2964,
null,
null,
1.9625,
null,
null,
0.0082,
null,
null,
0.1286,
null,
null
],
[
new Date(2007,0,28),
1.2956,
null,
null,
1.9707,
null,
null,
0.0082,
null,
null,
0.1285,
null,
null
],
[
new Date(2007,0,21),
1.2941,
null,
null,
1.9669,
null,
null,
0.0083,
null,
null,
0.1283,
null,
null
]
];
data.addColumn('date','Datum');
data.addColumn('number','EUR');
data.addColumn('string','Titel.EUR');
data.addColumn('string','Anmerkung.EUR');
data.addColumn('number','GBP');
data.addColumn('string','Titel.GBP');
data.addColumn('string','Anmerkung.GBP');
data.addColumn('number','YEN');
data.addColumn('string','Titel.YEN');
data.addColumn('string','Anmerkung.YEN');
data.addColumn('number','CHN');
data.addColumn('string','Titel.CHN');
data.addColumn('string','Anmerkung.CHN');
data.addRows(datajson);
return(data);
}
// jsDrawChart
function drawChartAnnotatedTimeLineIDf456e3929f() {
var data = gvisDataAnnotatedTimeLineIDf456e3929f();
var options = {};
options["width"] = 500;
options["height"] = 300;
options["displayAnnotations"] = true;
options["scaleColumns"] = [0,1];
options["scaleType"] = "allmaximized";
options["annotationsWidth"] = 10;
options["displayDateBarSeparator"] = false;
options["displayZoomButtons"] = false;
var chart = new google.visualization.AnnotatedTimeLine(
document.getElementById('AnnotatedTimeLineIDf456e3929f')
);
chart.draw(data,options);
}
// jsDisplayChart
function displayChartAnnotatedTimeLineIDf456e3929f()
{
google.load("visualization", "1", { packages:["annotatedtimeline"] });
google.setOnLoadCallback(drawChartAnnotatedTimeLineIDf456e3929f);
}
// jsChart
displayChartAnnotatedTimeLineIDf456e3929f()
<!-- jsFooter -->
//-->
</script></div>
<br />
<div id="AnnotatedTimeLineIDf456e3929f" style="height: 300px; width: 400px;">
</div>
<div>
R version 2.14.1 (2011-12-22) • <a href="http://code.google.com/p/google-motion-charts-with-r/">googleVis-0.2.13</a>
• <a href="http://code.google.com/apis/visualization/terms.html">Google Terms of Use</a> • <a href="http://code.google.com/apis/chart/interactive/docs/gallery/annotatedtimeline.html#Data_Policy">Data Policy</a><br />
<br />
<br />
Der Weg dahin ist gar nicht schwer. Zunächst müssen die Daten in die passende Form gebracht werden. Für obige Graphik heißt das:<br />
<ul>
<li>eine Spalte für das Datum</li>
<li>eine Spalte für die numerischen Werte, die hier als Linie angezeigt werden (die Wechselkurse, die angezeigt werden sollen)</li>
<li>eine Spalte für die Unterteilung der numerischen Werte in verschiedene Kategorien (hier: die verschiedenen Währungen)</li>
<li>eine Spalte für die Titel der Annotationen (die sowohl in, wie auch neben der Grafik angezeigt werden)</li>
<li>eine Spalte für die Annotationen</li>
</ul>
Das sieht dann folgendermaßen aus:<br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkwSpY45ymaUp7mTtqdyMJwfRghRVr6mlttG_bHsqRfGRWMDydBmDjXNHFiYXkkERBuYApfq6i0jsgOofE0rGgSYdwt5Oe-Qn6q-MxBQ_qxSbsmdiqtsxjuDmvMZN29B-TKoDZRSobXZY/s1600/daten.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="216" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkwSpY45ymaUp7mTtqdyMJwfRghRVr6mlttG_bHsqRfGRWMDydBmDjXNHFiYXkkERBuYApfq6i0jsgOofE0rGgSYdwt5Oe-Qn6q-MxBQ_qxSbsmdiqtsxjuDmvMZN29B-TKoDZRSobXZY/s320/daten.jpg" width="320" /></a></div>
<br />
Jetzt reichen zwei Befehle um die entsprechende Graphik zu erstellen und anzuzeigen:<br />
<br />
<span class="Apple-style-span" style="color: black; font-family: Monaco; font-size: 11px;">lineChart <span style="color: #0d2099;">=</span> <span style="color: #0d2099;">gvisAnnotatedTimeLine(</span>exchange.data<span style="color: #0d2099;">,</span> </span><br />
<span class="Apple-style-span" style="color: black;">
</span><br />
<div style="font: 11.0px Monaco; margin: 0.0px 0.0px 0.0px 4.0px; text-indent: -4.0px;">
<span class="Apple-style-span" style="color: black;"> datevar<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Datum"</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># Datumsvariable</span></span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
numvar<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Kurs"</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># numerische Werte</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
idvar<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Währung"</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># enthält die Gruppen</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
titlevar<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Titel"</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># enthält die Titel der Bemerkungen</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
annotationvar<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Anmerkung"</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># die Bemerkungen</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
options<span style="color: #0d2099;">=list(</span>displayAnnotations<span style="color: #0d2099;">=</span><span style="color: #c3892a;">TRUE</span><span style="color: #0d2099;">,</span><span style="color: #4f4f4f;">#Bemerkungen anzeigen</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
scaleColumns<span style="color: #0d2099;">=</span><span style="color: #af130b;">'[0,1]'</span><span style="color: #0d2099;">,</span> <span style="color: #4f4f4f;"># 2-Y-Achsen (eine links eine rechts)</span></div>
<div style="color: black; font: normal normal normal 11px/normal Monaco; margin-bottom: 0px; margin-left: 4px; margin-right: 0px; margin-top: 0px; text-indent: -4px;">
scaleType<span style="color: #0d2099;">=</span><span style="color: #af130b;">'allmaximized'</span><span style="color: #0d2099;">))</span></div>
<div>
<br />
erzeugt HTML-Code, der alle zum Anzeigen der Grafik benötigten Informationen enthält. Um das Ergebnis zu sehen benutzt man:</div>
<br />
<div style="font: 11.0px Monaco; margin: 0.0px 0.0px 0.0px 4.0px; text-indent: -4.0px;">
<span style="color: #0d2099;">plot(</span>lineChart<span style="color: #0d2099;">)</span></div>
<br />
. Im Browser öffnet sich dann die oben angezeigte Grafik. Das Kommando:<br />
<br />
<span class="Apple-style-span" style="font-family: Monaco; font-size: 11px;">lineChart</span><br />
<br />
zeigt bei Bedarf den der Grafik zugrundeliegenden HTML-Code an. Dieser enthält unter anderem auch die Daten selber. Somit kann man die Grafik einfach im Internet veröffentlichen indem man den HTML-Code auf dem eigenen Web-Space hoch lädt. Oder man kopiert den HTML-Code in den eigenen Blog.<br />
<br />
<b>Literatur:</b><br />
<ul>
<li><span class="Apple-style-span" style="-webkit-border-horizontal-spacing: 2px; -webkit-border-vertical-spacing: 2px; color: black; font-family: Verdana, 'DejaVu Sans', 'Lucida Grande', 'Bitstream Vera Sans', sans-serif; font-size: 13px;"><a href="http://journal.r-project.org/archive/2011-2/RJournal_2011-2_Gesmann+de~Castillo.pdf" style="color: #5a6488; outline-color: gray; outline-style: dotted; outline-width: thin; text-decoration: none;"><span class="Apple-style-span" style="color: black;"><em>Markus Gesmann and Diego de Castillo, </em></span>Using the Google Visualisation API with R </a> - R Journal (December 2011, Volume 3/2)</span></li>
</ul>
<br /></div>Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-80815481997513380822011-10-13T13:55:00.002+02:002011-10-14T12:35:51.487+02:00Datenimport mal anders<div style="text-align: justify;">
Wer kennt es nicht. Man liest ein Paper und denkt sich: Diese Daten hätte ich jetzt auch gern. Sei es um die Ergebnisse noch einmal selber nachzurechnen oder andere Methoden mit der vorgestellten zu vergleichen. <i></i></div>
<div style="text-align: justify;">
<i><br /></i></div>
<div style="text-align: justify;">
<i>[Einschub: Ok, es gibt wahrscheinlich den einen oder anderen, der diese Gefühl nicht kennt. Aber ganz abwegig ist es auch nicht :-)]</i></div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Glücklich, wer eine passende Grafik in dem Paper findet und das R-Paket <b>digitize</b> kennt.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Nehmen wir uns mal die folgende Grafik vor. Fehlen uns dazu die numerischen Werte (also die Koordinaten der einzelnen Punkte), so könnte man sich nun mit Lineal und Bleistift vor eine ausgedruckte Version setzen und die Werte der Punkte abmessen. Aber <b>digitize</b> macht es etwas bequemer.</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEigLzQoP_j9mMsjrGu-iXzR371lqKGZgGTbK4aVkTouZ97doSH-w2vnMfrA6822LN28ZNn3LPzk9ZBmhUh0-8KVk5BsK48ssvIVLDaDikKSgMEgdFTZRxwmTX4gMfurx9EuJGzdjpEEBJQ/s1600/Sample.jpeg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="326" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEigLzQoP_j9mMsjrGu-iXzR371lqKGZgGTbK4aVkTouZ97doSH-w2vnMfrA6822LN28ZNn3LPzk9ZBmhUh0-8KVk5BsK48ssvIVLDaDikKSgMEgdFTZRxwmTX4gMfurx9EuJGzdjpEEBJQ/s400/Sample.jpeg" width="400" /></a></div>
<div class="separator" style="clear: both; text-align: center;">
<br /></div>
Wir laden also das Paket <b>digitize</b> und müssen anschließend die Grafik in R importieren.<br />
<br />
<span style="font-family: "Courier New", Courier, monospace;">install.packages("digitize")</span><br />
<span style="font-family: "Courier New";">library(digitize)</span><br />
<span style="font-family: "Courier New";">axis = ReadAndCal("daten.jpg")</span><br />
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi2vhlyuXqfBX3YYw1YFG8FfnMwJsN0gi77KVE2L80tJ3o3dsi4PZBtjh5ziJ0KYAl752dMKdXuWHduTiJ1SdcYhD6eQNtrDhFcbcjHJKOuDd1q2TLQoA9Of4w5s7onEn1RjByAtzfTdtA/s1600/AchsenKalibrieren.jpeg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="325" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi2vhlyuXqfBX3YYw1YFG8FfnMwJsN0gi77KVE2L80tJ3o3dsi4PZBtjh5ziJ0KYAl752dMKdXuWHduTiJ1SdcYhD6eQNtrDhFcbcjHJKOuDd1q2TLQoA9Of4w5s7onEn1RjByAtzfTdtA/s400/AchsenKalibrieren.jpeg" width="400" /></a></div>
<br />
<div style="text-align: justify;">
Es öffnet sich gleich die importierte JPG-Datei. Zuerst müssen wir nun 4 Referenzpunkte setzen. Jeweils 2 auf der X-Achse und 2 auf der Y-Achse. Das geschieht durch einen Linksklick in der Grafik.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Im nächsten Schritt müssen die einzelnen Datenpunkte in der Grafik markiert werden. Das geschieht wieder mit einem Linksklick in die Grafik. <br />
<br />
<span style="font-family: "Courier New", Courier, monospace;">points = DigitData(col="red")</span><br />
<br />
Wer das ausprobiert, weiß spätestens jetzt warum sich das Paket nur für überschaubare Datensätze eignet. Manchmal dauert es etwas länger, aber wenn alle Punkte rot markiert sind beenden wir den Schritt indem man ESC drückt.</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjTOzDtUWGVf4LpEqr2v1ifFTDJIVCXSx7SBGQlY033elgaUB3RB_zIIc5dvRpKO40wZCMWKJ3Mdf8vO-NoXiGZ_SvZ4OE7Gu_hFaGieuiAr9LexcjyKzDF0VciLjf4FU_2zuwNXq6oqiQ/s1600/PunkteAuswa%25CC%2588hlen.jpeg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="325" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjTOzDtUWGVf4LpEqr2v1ifFTDJIVCXSx7SBGQlY033elgaUB3RB_zIIc5dvRpKO40wZCMWKJ3Mdf8vO-NoXiGZ_SvZ4OE7Gu_hFaGieuiAr9LexcjyKzDF0VciLjf4FU_2zuwNXq6oqiQ/s400/PunkteAuswa%25CC%2588hlen.jpeg" width="400" /></a></div>
<div style="text-align: justify;">
Zuletzt werden die gesammelten Achsen-Informationen benutzt um die Koordinaten der Punkte hochzurechnen. Dazu muss angeben werden an welchen Stellen die Punkte auf der X- bzw. Y-Achse gesetzt wurden (hier jeweils 0 und 2).</div>
<br />
<span style="font-family: "Courier New", Courier, monospace;">data = Calibrate(points, axis, 0, 2, 0, 2)</span><br />
<br />
Es ergeben sich die Folgenden gerundeten Werte:<br />
<br />
<div style="text-align: center;">
<table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse; text-align: center; width: 264px;"><colgroup><col span="3" style="width: 66pt;" width="88"></col></colgroup><tbody>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background: #6d7579; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; height: 12.75pt; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88"></td><td class="xl64" style="background: #6d7579; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88">X</td><td class="xl64" style="background: #6d7579; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88">Y</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">1</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">-2,5</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">-2,6</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">2</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1,2</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1,3</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">3</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0,3</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">4</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0,8</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0,4</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">5</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2,2</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">6</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0,1</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">0,3</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">7</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">-0,3</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">-0,3</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">8</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2,3</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2,8</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">9</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">?</td><td class="xl64" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">?</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl64" height="17" style="background-color: transparent; border-bottom: #6d7579 0.5pt solid; border-left: #6d7579 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">10</td><td class="xl64" style="background-color: transparent; border-bottom: #6d7579 0.5pt solid; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2,0</td><td class="xl64" style="background-color: transparent; border-bottom: #6d7579 0.5pt solid; border-left: #f0f0f0; border-right: #6d7579 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">2,1</td></tr>
</tbody></table>
</div>
<div style="text-align: center;">
<br /></div>
<div style="text-align: center;">
<br /></div>
<div style="text-align: justify;">
Abgesehen von minimalen Abweichungen stimmen diese Daten mit denen die ich tatsächlich benutzt habe um die Grafik zu erzeugen überein. Allein Punkt 9, der von den Koordinaten identisch zu Punkt 3 ist kann mit der Methode natürlich nicht gefunden werden. 2 direkt übereinanderliegende Punkte können in der Grafik schließlich nicht auseinandergehalten werden.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
<b>Literatur:</b></div>
<div style="text-align: justify;">
<ul>
<li><span class="Apple-style-span" style="font-family: inherit;"><span class="Apple-style-span" style="font-family: inherit; font-weight: bold;">Timothée Poisot: </span><i>The digitize Package: Extracting Numerical Data from Scatterplots</i> - R Journal Vol 3/1 (June 2011). </span>http://journal.r-project.org/archive/2011-1/RJournal_2011-1_Poisot.pdf (Stand: 10. Oktober 2011).</li>
</ul>
</div>
<div align="justify" style="text-align: center;">
</div>Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-49850658375839649262011-10-07T14:12:00.002+02:002011-10-07T14:17:22.751+02:00(Welt)-Karten mit JMP<div style="text-align: justify;">
In Anlehnung an den letzten Beitrag möchte ich noch erwähnen, dass R natürlich nicht die einzige Software ist, die sich zum Erstellen von Karten eignet. Hier möchte ich noch kurz anhand der FIFA-Damen-Weltrangliste zeigen, wie sich die gleichen Grafiken in JMP erzeugen lassen.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Die benutzten Daten finden sich wieder im <a href="http://www.fifa.com/worldranking/rankingtable/women/index.html">Internet</a>.</div>
<div style="text-align: justify;">
<br /></div>
<div style="text-align: justify;">
Nachdem die Daten in JMP eingelesen sind erzeugen wir die Karte nun im Graph Builder (Graph ⇒ Graphik erstellen).</div>
<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjBw312GxTK3UXYb9nZQwcSsDEb6pJ2pFdCy8aAXhGGPwirdDeDRWucZPmfvL9atXrcWzhs_aCc3lACmQ0iT69IrpmdVmJyol7JG0Zhiq1g__oz07M2jazQGd0HrXPxT5qfyFSzENfQFvY/s1600/Graph+Builder.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="229" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjBw312GxTK3UXYb9nZQwcSsDEb6pJ2pFdCy8aAXhGGPwirdDeDRWucZPmfvL9atXrcWzhs_aCc3lACmQ0iT69IrpmdVmJyol7JG0Zhiq1g__oz07M2jazQGd0HrXPxT5qfyFSzENfQFvY/s400/Graph+Builder.png" width="400" /></a></div>
Zieht man die Ländernamen in den Shape- (bzw. Form-) Bereich des Graphbuilders entsteht automatische die entsprechende Weltkarte.<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEil4trAIRPgDXicVOwtNjeOE-1-YABj7u40pT_KOfqNBOU10mvoxoxWW4lHX1vS7M2Tjgq60iOJBiEbz5oT2KjrPXpLm2dOZi-KUwWn6Z6wzjg8E_Mx7xYxFbns9RHwvBIgJ-7mLJG0aiI/s1600/graph+builder+2.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="241" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEil4trAIRPgDXicVOwtNjeOE-1-YABj7u40pT_KOfqNBOU10mvoxoxWW4lHX1vS7M2Tjgq60iOJBiEbz5oT2KjrPXpLm2dOZi-KUwWn6Z6wzjg8E_Mx7xYxFbns9RHwvBIgJ-7mLJG0aiI/s400/graph+builder+2.png" width="400" /></a></div>
Danach fehlen noch die Farben gemäß der Punktezahl der verschiedenen Teams. Dazu zieht man die Variable Pts auf die Karte.<br />
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgnnOruLV5FevJQqSXLVLODIoyYO7OE2LzRFdLh0LcDBNp473hAiITGscMSav34MTSanbTgJSKuLE83cOw18-zIHW1dcEwZNoK-AhG35WAjh7yICf_Qonx6pPF24yOlcXAC78cYddbq574/s1600/graph+builder+3.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="241" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgnnOruLV5FevJQqSXLVLODIoyYO7OE2LzRFdLh0LcDBNp473hAiITGscMSav34MTSanbTgJSKuLE83cOw18-zIHW1dcEwZNoK-AhG35WAjh7yICf_Qonx6pPF24yOlcXAC78cYddbq574/s400/graph+builder+3.png" width="400" /></a></div>
<div style="text-align: justify;">
Nach einem bisschen Heraumspielen an den Farbverläufen (Rechtsklick auf die Legende ⇒ Verlauf) kann man eine Karte erzeugen, die der mit R erstellten verblüffend ähnlich sieht. Abgesehen davon, dass wir hier den Damen-Fußball und mit R den Herren-Fußball analysiert haben.</div>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQBehmAtTlJtmGcc4h8gl0Wge4uW6Iy9z8BIfl7oWyz3w3t2tOSD2L0v5FD207BjhJonF936GQKlgAIVp6ZhYzSGdzrP5u_Wsal_140abFMo0UmAR6xRehGrIFYZwym_rf4LP2ha9X4AU/s1600/Final.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" height="257" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQBehmAtTlJtmGcc4h8gl0Wge4uW6Iy9z8BIfl7oWyz3w3t2tOSD2L0v5FD207BjhJonF936GQKlgAIVp6ZhYzSGdzrP5u_Wsal_140abFMo0UmAR6xRehGrIFYZwym_rf4LP2ha9X4AU/s400/Final.png" width="400" /></a></div>
Auch hier stimmen die Namen der Länder nicht überall mit den von JMP erwarteten Namen überein. Das sieht man an den weißen Flächen auf der Karte (Zentral-Afrika, China,...).<br />
<br />
<b>Literatur:</b><br />
<ul>
<li><b>FIFA Weltrangliste der Damen </b>(Stand: 7. Oktober 2011) - http://www.fifa.com/worldranking/rankingtable/women/index.html</li>
</ul>
Unknownnoreply@blogger.com0tag:blogger.com,1999:blog-276389194507638774.post-53972135378453683742011-10-07T12:10:00.000+02:002011-10-14T12:37:25.398+02:00Weltkarten mit R<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9X-ph7uXgsFyBSP8FliIp05uNLSurBw-NO_0VNKzKxLD_4pZ55Fe0WM4DRyjNwojA1RoSXmVcbK-Pl3FhsHUB4B9c2I7XvE1yKpYS1ShZcoi3AggOINYpc58UkjxC6vClD39ChpZPLLk/s1600/feine+Version.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><span class="Apple-style-span" style="font-family: inherit;"></span></a></div>
<div style="text-align: justify;">
In der neuesten Ausgabe des <a href="http://journal.r-project.org/">R Journals</a> findet sich ein lesenswerter Artikel zu dem Packet <b>rworldmap</b>. Diese erlaubt es mit wenigen Befehlen eine Weltkarte zu erstellen, in der die einzelnen Länder individuell eingefärbt sind. Wie einfach das ist will ich am Beispiel der FIFA-Männer-Fußball-Weltrangliste zeigen.</div>
<div style="font: 12px Helvetica; margin: 0px; min-height: 14px; text-align: justify;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<table cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9X-ph7uXgsFyBSP8FliIp05uNLSurBw-NO_0VNKzKxLD_4pZ55Fe0WM4DRyjNwojA1RoSXmVcbK-Pl3FhsHUB4B9c2I7XvE1yKpYS1ShZcoi3AggOINYpc58UkjxC6vClD39ChpZPLLk/s1600/feine+Version.png" imageanchor="1" style="clear: right; margin-bottom: 1em; margin-left: auto; margin-right: auto;"><img border="0" height="326" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9X-ph7uXgsFyBSP8FliIp05uNLSurBw-NO_0VNKzKxLD_4pZ55Fe0WM4DRyjNwojA1RoSXmVcbK-Pl3FhsHUB4B9c2I7XvE1yKpYS1ShZcoi3AggOINYpc58UkjxC6vClD39ChpZPLLk/s400/feine+Version.png" width="400" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Länder eingefärbt nach ihren Punkten auf der Männer-FIFA-Weltrangliste</td></tr>
</tbody></table>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Zunächst müssen einige Pakete installiert werden. Wer sich mehr mit Landkarten in R beschäftigen möchte sollte einen Blick auf die Liste der mitinstallierten Pakete werfen: sp, maptools beschäftigen sich ebenfalls mit diesen Themen. </span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"></span></div>
<div style="color: #0d2099; font: 11px Monaco; margin: 0px 0px 0px 4px; text-align: left; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><span class="Apple-style-span" style="font-family: inherit; font-size: small;">install.packages(<span style="color: #af130b;">"rworldmap"</span>)</span></span></div>
<div style="color: #0d2099; font: 11px Monaco; margin: 0px 0px 0px 4px; text-align: left; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><span class="Apple-style-span" style="font-family: inherit; font-size: small;">library(<span style="color: black;">rworldmap</span>)</span></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Nachdem die Bibliothek geladen ist müssen die Daten beschafft werden. Für unser Beispiel finden sich die Daten im <a href="http://www.fifa.com/worldranking/rankingtable/index.html">Internet</a>. Etwas aufbereitet sehen die Daten dann etwa so aus:</span></div>
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"></span><br />
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: center; text-indent: -4px;">
<table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse; width: 264px;"><colgroup><col style="width: 66pt;" width="88"><col style="width: 66pt;" width="88"><col style="width: 66pt;" width="88"></colgroup><tbody>
<tr height="17" style="height: 12.75pt;"><td class="xl71" height="17" style="background: #6d7579; border-bottom: #f0f0f0; border-left: windowtext 1pt solid; border-right: #f0f0f0; border-top: windowtext 1pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; height: 12.75pt; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88">Rank</td><td class="xl72" style="background: #6d7579; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: windowtext 1pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88">Team</td><td class="xl73" style="background: #6d7579; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 1pt solid; border-top: windowtext 1pt solid; color: white; font-family: Arial; font-size: 10pt; font-weight: 700; mso-pattern: #6D7579 none; text-decoration: none; text-line-through: none; text-underline-style: none; width: 66pt;" width="88">Points</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl66" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: windowtext 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">1</td><td class="xl67" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: windowtext 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Spain</td><td class="xl68" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: windowtext 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1605</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl69" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">2</td><td class="xl63" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Netherlands</td><td class="xl70" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1571</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl69" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">3</td><td class="xl63" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Germany</td><td class="xl70" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1290</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl69" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">4</td><td class="xl63" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Uruguay</td><td class="xl70" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1184</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl69" height="17" style="background-color: transparent; border-bottom: #f0f0f0; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">5</td><td class="xl63" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Portugal</td><td class="xl70" style="background-color: transparent; border-bottom: #f0f0f0; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1158</td></tr>
<tr height="17" style="height: 12.75pt;"><td class="xl69" height="17" style="background-color: transparent; border-bottom: windowtext 0.5pt solid; border-left: windowtext 0.5pt solid; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; height: 12.75pt; text-decoration: none; text-line-through: none; text-underline-style: none;">6</td><td class="xl63" style="background-color: transparent; border-bottom: windowtext 0.5pt solid; border-left: #f0f0f0; border-right: #f0f0f0; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">Italy</td><td class="xl70" style="background-color: transparent; border-bottom: windowtext 0.5pt solid; border-left: #f0f0f0; border-right: windowtext 0.5pt solid; border-top: #6d7579 0.5pt solid; color: black; font-family: Arial; font-size: 10pt; font-weight: 400; text-decoration: none; text-line-through: none; text-underline-style: none;">1142</td></tr>
</tbody></table>
</div>
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"></span><br />
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<br /></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
</div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Wenn die Daten in R sind, brauchen wir nur noch 3 Befehle. Der erste verbindet die Information der Daten mit einer Weltkarte. Der zweite erstellt eine neue leere Weltkarte und der dritte benutzt die zuvor erzeugte Information um die leere Weltkarte einzufärben.</span><br />
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /> </span></div>
<div align="justify" style="font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">fifa.map.info <span style="color: #0d2099;">=</span> <span style="color: #0d2099;">joinCountryData2Map(</span>fifa.ranking<span style="color: #0d2099;">,</span> joinCode<span style="color: #0d2099;">=</span><span style="color: #af130b;">"NAME"</span><span style="color: #0d2099;">, </span>nameJoinColumn<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Team"</span><span style="color: #0d2099;">)</span></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="font-family: Helvetica; font-style: normal; font-variant: normal; font-weight: normal; line-height: normal; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: "Courier New", Courier, monospace;">150 codes from your data successfully matched countries in the map</span></div>
<div style="font-family: Helvetica; font-style: normal; font-variant: normal; font-weight: normal; line-height: normal; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: "Courier New", Courier, monospace;">57 codes from your data failed to match with a country code in the map</span></div>
<div style="font-family: Helvetica; font-style: normal; font-variant: normal; font-weight: normal; line-height: normal; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: "Courier New", Courier, monospace;">96 codes from the map weren't represented in your data</span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Der Output sagt uns, dass 150 der Länder richtig erkannt wurden, 57 nicht erkannt wurden und 96 Länder in der Karte sind, die nicht von uns angesprochen wurden. Nicht erkennen kann R Länder die "falsch" geschrieben sind. So ist es natürlich ein Unterschied ob in den Daten USA (was bei uns der Fall ist) oder United States (was von R bzw. vom Packet <b>rworldmap</b> erwartet wird) steht. Im Moment wollen wir diese Problematik kurzfristig beiseite lassen und einfach die Karte zeichnen.</span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="color: #0d2099; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">mapDevice()</span></div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><span style="color: #0d2099;">mapCountryData(</span>fifa.map.info<span style="color: #0d2099;">,</span> nameColumnToPlot<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Points"</span><span style="color: #0d2099;">)</span></span></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><img border="0" height="327" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjyrp_W4qXn5DvvMA1jFEbp0ZoxVyn43e3CQlXeT8em1HS0mu_7QqsS1W2am2vhNlr8Hzim-vkCKOdN6RwzXlU-vkGlaVK-Pb6El5z4NFfDAahjR1giKVWYInqtvF2mE6Yp-gel-aymzg4/s400/La%25CC%2588nder+roh.png" style="margin-left: auto; margin-right: auto;" width="400" /></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Die weißen Länder wurden nicht erkannt</td></tr>
</tbody></table>
<div class="separator" style="clear: both; text-align: center;">
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjyrp_W4qXn5DvvMA1jFEbp0ZoxVyn43e3CQlXeT8em1HS0mu_7QqsS1W2am2vhNlr8Hzim-vkCKOdN6RwzXlU-vkGlaVK-Pb6El5z4NFfDAahjR1giKVWYInqtvF2mE6Yp-gel-aymzg4/s1600/La%25CC%2588nder+roh.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><span class="Apple-style-span" style="font-family: inherit;"></span></a></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-indent: -4px;">
<span class="Apple-style-span" style="font-size: small;">Es ergibt sich die eingefärbte Karte mit einigen weißen Flecken. Dies sind Länder, die im ersten Schritt nicht erkannt wurden. Um diese Problem zu beheben müssen die Ländernamen im Datensatz an die Ländernamen, die das Packet kennt anpassen. Dabei hilft der Befehl</span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="color: #0d2099; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">identifyCountries()</span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Mit Mausklick auf die Karte werden die unerkannten Ländernamen identifiziert. Diese können dann anschließend im Datensatz korrigiert werden.</span></div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; min-height: 14px; text-align: justify; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span></div>
<table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhVHzkRj7Pkq8U1AuYydjOGBNN1f5UKD46kS3fgdw18WgZ9vqAJhTXKQLKpFIWwHC6UWhDmYyv3QRFPUBT3PGXTJ0KgrK9B2ZDX2kYscCXnr5dh_f1mBXYooC-ojBCnEl1ojW6CumUK0tI/s1600/La%25CC%2588nder.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" height="327" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhVHzkRj7Pkq8U1AuYydjOGBNN1f5UKD46kS3fgdw18WgZ9vqAJhTXKQLKpFIWwHC6UWhDmYyv3QRFPUBT3PGXTJ0KgrK9B2ZDX2kYscCXnr5dh_f1mBXYooC-ojBCnEl1ojW6CumUK0tI/s400/La%25CC%2588nder.png" width="400" /></a></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Auf der Suche nach den richtigen Ländernamen</td></tr>
</tbody></table>
<div class="separator" style="clear: both; text-align: center;">
<span class="Apple-style-span" style="font-family: inherit; margin-left: 1em; margin-right: 1em;"></span></div>
<div class="separator" style="clear: both; text-align: center;">
<span class="Apple-style-span" style="font-family: Helvetica;">Nachdem die entsprechenden Ländernamen korrigiert wurden ergibt sich die endgültige Karte.</span></div>
<table cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto; text-align: center;"><tbody>
<tr><td style="text-align: center;"><img border="0" height="327" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjBuPHYUJBXQvOWmqj_E0vVHthje5ax_-DWugZfU8zr-w07nHzaqptSHJmT-ibpmHeBufkTN7zbEzGiwbmJN69DNtLa9f2_afN0STaPeEEyM1z1jmEjrEwU-X_fp_JFIWvudZe_GeoM7So/s400/final.png" style="margin-left: auto; margin-right: auto;" width="400" /></td></tr>
<tr><td class="tr-caption" style="text-align: center;">Mit korrigierten Ländernamen<br />
<div style="text-align: justify;">
<span class="Apple-style-span" style="font-size: small;"><br /></span></div>
</td></tr>
</tbody></table>
<div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;">Das Skript, dass die oberste Karte erzeugt verwendet noch einige zusätzliche Optionen. Der zugehörige Code findet im Anhang.</span><br />
<div style="text-align: justify;">
<span class="Apple-style-span" style="font-family: inherit; font-size: small;"><br /></span><br />
<b><span class="Apple-style-span" style="font-family: inherit; font-size: small;">Literatur:</span></b></div>
</div>
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-indent: -4px;">
</div>
<ul>
<li><span class="Apple-style-span" style="font-family: inherit;"><b>Andy South:</b><i> rworldmap: A New R package for Mapping Global Data</i> - The R Journal Volume 3/1, June 2011</span></li>
<li><span class="Apple-style-span" style="font-family: inherit;"><b>Fifa Weltrangliste der Männer</b> (Stand: 7. Oktober 2011) - http://www.fifa.com/worldranking/rankingtable/index.html</span></li>
</ul>
<br />
<div>
<b>Anhang:</b></div>
</div>
<div>
<br /></div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
# Erstellen einer Karte mit der FIFA Weltrangliste</div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
# Autor: Sebastian Hoffmeister - Statcon</div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
# 7.10.2011</div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; min-height: 15px; text-indent: -4px;">
<br /></div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span style="color: #0d2099;">library(</span><span style="color: black;">rworldmap</span><span style="color: #0d2099;">)</span><span style="color: black;"> </span># Laden der benötigten Pakete</div>
<div style="color: #0d2099; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
library(<span style="color: black;">classInt</span>)</div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; min-height: 15px; text-indent: -4px;">
<br /></div>
<div style="color: #af130b; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span style="color: black;">fifa.ranking </span><span style="color: #0d2099;">=</span><span style="color: black;"> </span><span style="color: #0d2099;">read.csv(</span>".../Fifa Ranking.csv"<span style="color: #0d2099;">)</span><span style="color: black;"> </span><span style="color: #4f4f4f;"># Einlesen der Daten</span></div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
fifa.map.info <span style="color: #0d2099;">=</span> <span style="color: #0d2099;">joinCountryData2Map(</span>fifa.ranking<span style="color: #0d2099;">,</span> joinCode<span style="color: #0d2099;">=</span><span style="color: #af130b;">"NAME"</span><span style="color: #0d2099;">,</span> nameJoinColumn<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Team"</span><span style="color: #0d2099;">)</span> <span style="color: #4f4f4f;"># Daten mit der Karte verbinden</span></div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; min-height: 15px; text-indent: -4px;">
<br /></div>
<div style="color: #af130b; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span style="color: black;">palette </span><span style="color: #0d2099;">=</span><span style="color: black;"> </span><span style="color: #0d2099;">colorRampPalette(c(</span>"brown4"<span style="color: #0d2099;">,</span><span style="color: black;"> </span>"darkgoldenrod1"<span style="color: #0d2099;">,</span><span style="color: black;"> </span>"darkolivegreen4"<span style="color: #0d2099;">))(</span><span style="color: #065118;">80</span><span style="color: #0d2099;">)</span><span style="color: black;"> </span><span style="color: #4f4f4f;"># Farbverlauf selbst definieren.</span></div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span style="color: black;">classInt </span><span style="color: #0d2099;">=</span><span style="color: black;"> </span><span style="color: #0d2099;">classIntervals(</span><span style="color: black;">fifa.map.info</span><span style="color: #0d2099;">[[</span><span style="color: #af130b;">"Points"</span><span style="color: #0d2099;">]],</span><span style="color: black;"> n</span><span style="color: #0d2099;">=</span><span style="color: #065118;">80</span><span style="color: #0d2099;">,</span><span style="color: black;"> style</span><span style="color: #0d2099;">=</span><span style="color: #af130b;">"jenks"</span><span style="color: #0d2099;">)</span><span style="color: black;"> </span># Eigene anzahl an Farb-Intervallen definieren</div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
catMethod <span style="color: #0d2099;">=</span> classInt<span style="color: #0d2099;">[[</span><span style="color: #af130b;">"brks"</span><span style="color: #0d2099;">]]</span></div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; min-height: 15px; text-indent: -4px;">
<br /></div>
<div style="color: #4f4f4f; font: 11px Monaco; margin: 0px 0px 0px 4px; text-indent: -4px;">
<span style="color: #0d2099;">mapDevice()</span><span style="color: black;"> </span># Die leere Karte erstellen</div>
<div style="font: 11px Monaco; margin: 0px 0px 0px 4px; text-align: justify; text-indent: -4px;">
<span style="color: #0d2099;">mapCountryData(</span>fifa.map.info<span style="color: #0d2099;">,</span> nameColumnToPlot<span style="color: #0d2099;">=</span><span style="color: #af130b;">"Points"</span><span style="color: #0d2099;">,</span> colourPalette <span style="color: #0d2099;">=</span> palette<span style="color: #0d2099;">,</span> catMethod<span style="color: #0d2099;">=</span>catMethod<span style="color: #0d2099;">)</span> <span style="color: #4f4f4f;"># Die leere Karte befüllen</span></div>
<br />
<div style="font: 12px Helvetica; margin: 0px 0px 0px 4px; text-align: center; text-indent: -4px;">
<br /></div>Unknownnoreply@blogger.com0