Finnish Parliamentary Election, 2011 - Data Mining Visualization

The distribution of the values and opinions of the Finnish voters (surface), political parties (abbreviation) and the presidents of each party (name) in the new Parliament are visualized. Two databases were combined: the voting advice application data provided by Helsingin Sanomat and the results of the parliamentary elections 2011. The level of surface is derived by multiplying each candidate opinion distribution matrix by the number of votes. It is assumed that in average citizens voted politician with the same opinions.

Data was reduced to two principal components (PCA). The first principal component has as high a variance as possible and the next component the second highest, etc.
In this visualization, left vs. right wing classification can be detected. Also parties with liberal vs. conservative opinions are classified different groups. The value grid is produced with a mathematical method without explicitly using the questions of the voting advice application.

Analysis examples:
Timo Soini and Anni Sinnemäki have opposite opinions and these politicians have more 'radical' opinions than the median of candidates in their party. Jyrki Katainen (KOK), Mari Kiviniemi (KESK) and Stefan Wallin (RKP) have similar values. Jutta Urpilainen did not answers to the voting advice application, so her opinions were approximated by the candidates' opinions of the SDP party. Therefore her name and party tag are in the same place. The median opinions of candidates for parties who did not get seats in the new Parliament are visualized too, see list of political parties in Finland [http://www.vaalit.fi/15485.htm].

Category:

Education

Tags:

• Parliamentary election
• Finland
• PCA
• visualization