Jigsawites use Text Mining to Analyze Election Manifestos

Just before Elections, we definitely had a lot of Questions in mind. The most obvious question was which political party was going to win? Two Jigsawites, Sakshi Batra and Sameer Bhatia, under the mentorship of Abhirami Sankar tried to answer this question and took it up as a project. They decided to thoroughly analyse the various goals of the political parties (from the manifestos). Let’s hear from Abhirami how they approached the project and the analysis they came up with.


As we got more into the task at hand, more questions were brimming. How different are the manifesto’s of each political party? Which sector was Congress and BJP more focused on? What was so different about AAP’s goals (being a relatively new party?) Was the BJP manifesto more elaborate in terms of their goals? What unique features did each party focus on? What promises have these parties made to the people of India?

It definitely made a lot of sense to use analytics to understand the manifesto’s instead of manually traversing through the pages. This is where text mining helped us! Using Text Mining techniques, we did a very interesting analysis to understand the political manifestos of Elections Data – INC, BJP, AAP and CPI (M)

Text mining refers to the process of deriving high-quality information from text. This is basically turning text into numbers

Is just understanding the documents enough? What about real time sentiments of the people towards BJP, Congress, AAP, etc?This is where Social Media analysis helped us –  We also tracked the sentiments of the parties from different social media sources (like twitter, Facebook ) for a given period of time.

We found out post elections that the clear winner was BJP.This is because their manifesto was very elaborate and also through social media analysis we found out that BJP had a very high share of Positive sentiments on twitter and Facebook.

Let me now elaborate on some of the various techniques we used:
1. Text Mining Techniques Used in R

Text mining techniques on the given manifesto’s helped us understand what each party is trying to convey to us. From this we were able to analyse what were the top priorities of each party for the election year .Following is a list of the techniques used:

  • Word Cloud: Visualisation of the most frequently occurring words in the given manifesto.

  • Frequency Analysis:

    To understand the proportion of words in a given manifesto since we see that all these parties focus on similar high level goals but the proportion is different

  • We can clearly see that Congress is trying to appease the women, youth and the minorities

  •  BJP is focusing on making use of technology to solve problems, introducing many educational reforms and development of infrastructure
  • AAP‘s focus is on Media, Security ,Police and ways to reduce corruption


  • Network of words:  
  • This is useful for identifying clusters of words which are more co-related to each other.

After doing a comparison amongst parties, Congress had  distinct network of nodes as compared to BJP and AAP. For Congress network:

  • The government will “ensure” is one network which is linked to the economic and social growth network .It is also linked to financial and infrastructure development.
  • We have another network which tells us that congress talks about secularism, communalism, and narrow-mindedness. This is because Congress claims BJP to be secular and narrow minded and Congress is secular in its approach.
  • The next network is where we see a mention of north, east and strife. This is because Congress claims to place a special focus on ensuring educational access to the north civil strife effected areas and the north-east region.
  • Women and children are associated with the welfare

2. Social media Analysis:

We used social media tools to try and understand the sentiments of the people towards BJP, Congress and AAP. The sentiments were analysed for the keywords ”congress”,”aap” and “bjp” across various social media sources such as Facebook, twitter, google + etc. for a period of 1 week(April 25th to May 5 th)

This project was an enjoyable and great learning experience for all of us. Thanks Sakshi Batra and Sameer Bhatia, for your hard work. I am sure you took back a lot from this experience. As Jigsaw mentors working with students on such projects is one of the most interesting facets of our job. We are actually able to show our students how to use what they learn to analyse real life problems. That’s a great feeling of accomplishment for both the student and the teacher.

 Related Articles:
Jigsaw Academy Poll Twitter Analysis Splashed Across the Media
DNA Quotes Jigsaw’s Social Media Analysis
Indian Elections 2014: Jigsaw Faculty Analyses the Parties Twitter Campaigns
Jigsaw Students ask the Question – What would happen if the World Cup was a League?

Interested in a career in Data Science?
To learn more about Jigsaw’s Data Science with SAS Course – click here.
To learn more about Jigsaw’s Data Science with R Course – click here.
To learn more about Jigsaw’s Big Data Course – click here.


Related Articles

} }
Request Callback