PollAnalytics for Elections - Criminal Cases, Education and Assets

This post is the outcome of the data analysis that we did of wining candidates focusing on criminal cases against each along with their educational qualification and assets they have.

The findings, for now, are solely of the 2009 Lok Sabha elections, but in our future posts we will share a full time line of 3-4 elections, if comprehensive data is available.

We start off with a bar-chart that is a combination of all three points-of-interest, i.e. Criminal Cases, Education and Assets. It gives a very fair picture as to the type of candidates who were given tickets to contest in the election.

The graph above is a visualization of winning candidates (Lok Sabha 2009) that have most criminal cases based on their asset class and education. Two major pointers that may be observed here are:

• Candidates who are graduates and belong to the “High” asset class, have the max number of criminal cases against them
• Candidates belonging to the “Low” asset class, and having completed their education only till 10^th standard, have the max number of cases against them. The same goes with “Medium” asset class candidates

Winning candidates who are 10^th and 12^th pass along with graduates together constitute 58% of all the criminal cases lodged. Thus, "educated" candidates have the most criminal cases against them.

The High Asset class group "contributes" the most to the criminal cases lodged. Evidently more money means more power which in turn helps the candidate to secure a ticket and eventually the seat to political power.

The above chart tells about the conversion rate of candidates i.e. what percentage of candidates in which education group was able to win the election. Far ahead from the rest, 22% of all graduate professionals that contested in the elections won in their respective constituencies.

Furthermore using data particularly of Karnataka and Andhra Pradesh, a predictive model was applied on it. Based on attributes like Education, Criminal cases and Assets, a total of 42 instances out of 45 were predicted correctly, as to whether that candidate will Win or Lose.