[DEPRECATED] Python script to analyze correlations in the wahl-o-mat data
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


[DEPRECATED] correl-o-mat

Deprecation notice

This script no longer works, because the offline version (zip file) of Wahl-O-Mat is no longer available and the https://www.bpb.de has no interest in providing the data in machine readable format. However some smart and hard-working people have found workarounds by parsing the data out from the PDFs provided by the bpb:


correl-o-mat is a python script for analyzing correlations from the data of the wahl-o-mat.

It generates a table with all the answers of all parties for every question in the wahl-o-mat. In addition, it generates a correlation matrix counting all questions where the parties agree with each other. This shows, which parties share their opinions and therefor are similar to each other. Thereby the correlation matrix also suggests which parties could form coalitions and which parties better do not because they are too different from each other.

System Requirements

Tested on Mac OS X 10.8.4 with Python 2.7.3.. However the script should also run on Linux or Windows.


  1. Download the wahlomat.zip file for an election, e.g. for Bayern 2013 or from the wahl-o-mat archive.

  2. Run the correlomat script giving the path to wahlomat.zip file as the only argument

python correlomat.py wahlomat.zip

This will generate two files in the working directory of the script. The file table.csv holds the answers of all parties to all questions. The file correlation.csv contains the correlation matrix of all parties, showing the number of the questions each party agrees on with the other parties.