we may have used their api

Inspiration

Our idea came up in a conversation when we were wondering how we can identify an individual’s political preferences in social media. It seems quite obvious that we interact with content and people of similar views, but it's also the case that the environment we're in exerts pressure on ourselves. Hence our motivation to build a model and investigate what happens in the real world.

What it does

It shows the evolution of a social network architecture on a subset of Twitter users from Florida. We investigate the change in their political engagement as well as changes in political affiliation clusters.

How We built it

We built it by scraping a sample of social network connections in the state of Florida. Then, we applied an algorithm to identify political engagement of the most central nodes in that network. From there, we populated the engagement metrics across all users through the connections to finally apply bisection algorithms showing us the political clusters.

Challenges We ran into

The main challenge we faced was to balance out the network representation so that it's not skewed towards any political cluster. We did so by extracting random starting set of users and then analysing their connections deeper. This way, we ensured that we also did not impose any specific political engagement on our sample.

Accomplishments that We're proud of

We are especially proud of the fact that from this analysis we can infer actual changes in the political engagement and cluster structure within the last 2 months. This shows the validity of the approach and an uncovered potential of Social Network Analysis that could be used to liquidate the political life in a similar manner to how financial market have been liquidated in the last half a century.