Miscellaneous‎ > ‎

Twitter Adoption in Indian Politics

This project examines patterns of twitter adoption and activity by Indian politicians and the factors that influence them. There are persistent differences in the adoption of internet and communication technologies in India along socioeconomic lines. This research hopes to determine if similar digital divides exist in the political community and their use of micro-blogging services such as Twitter. We also examine if the makeup of their electoral constituencies play a role in the adoption of such services.

We further analyze political rivalries and loyalties on twitter networks, specifically if the bandwagon effect explains why politicians are joining twitter – i.e. a result of their offline networks such as party membership. However, proving the existence of such a bandwagon effect is hard. We attempt to do through an analysis of the dates of joining and a similarity analysis of tweets/followers, to see if party membership/rivalries in the respective states play any roles in both the politicians joining Twitter as well as their Twitter activity. (Twitter activity, in our project, is measured by the frequency and content of tweets.) This will also enable us to identify if loyalty in the political sphere in India is limited to a few influential individuals or is to the political parties.

Collection of Initial Data
The Indian Lower House (Lok Sabha) and Indian Upper House (Rajya Sabha) website was parsed for information with regards to sitting MPs. The information collected included -

  1. Name of MP
  2. Parliamentary Constituency (for LS MP)
  3. Political Party
  4. State
  5. Date of Birth
  6. Education Qualifications – This was coded as two variables – a) Level of Education b) Quality of Education
    • Level of Education is ranked as per the following : 1 – School + Others 2 – Bachelors Degree 3- Masters Degree 4 – PhD degree
    • Quality of Education is ranked as per the following classifications : International Universities > Institutes of National Importance > National and State Universities > Local Colleges > School + Others

From the Association for Democratic Reforms, we obtained the asset records for these politicians. While these records are available on the Parliament website, ADR has consolidated the data and made it freely available on their website.

Collection of Constituency Data

Data for each assembly constituency (AC) is estimated from the latest 2011 census data. As the 2011 Census website at the moment only provides data at the district level, each AC within a district is assumed to have characteristics similar to the parent district. This is then averaged across multiple ACs to provide us with parliamentary constituency estimates. AC to PC mapping has been provided by Indiavotes.in, which is also sharing consolidated electoral data.

The data we are using to judge the penetration of the internet and socioeconomic characteristics of the constituency are :

  • Availability of assets: Computer/Laptop, Per cent
  • Availability of assets: Computer/Laptop – with internet – Households, Per cent
  • Availability of assets: Computer/Laptop – without internet – Households, Per cent
  • Urban / Rural ratio
  • Literacy level

Extending the Database

Extending Database to Rajya Sabha and State Legislative Assemblies, along with non-sitting politicians/parties is accomplished using Amazon Turk.

Logistic Regression

Downloading Twitter Data

Python code

Visualizing Twitter Data
Automated dashboards that use database entries to extract data through Twitter API. Processed data is displayed (using D3 Javascript library) ()

Data Mining