Using sentiment analysis to define twitter political users' classes and their homophily during the 2016 American presidential election

This paper proposes an analysis of political homophily among Twitter users during the 2016 American Presidential Election. We collected 4.9 million tweets of 18,450 users and their contact network from August 2016 to November 2016. We defined six user classes regarding their sentiment towards Donald Trump and Hillary Clinton: whatever, Trump supporter, Hillary supporter, positive, neutral, and negative. Next, we analyzed their political homophily in three scenarios. Firstly, we analyzed the Twitter follow, mention and retweet connections either unidirectional and reciprocal. In the second scenario, we analyzed multiplex connections, and in the third one, we analyzed friendships with similar speeches. Our results showed that negative users, users supporting Trump, and users supporting Hillary had homophily in all analyzed scenarios. We also found out that the homophily level increase when there are reciprocal connections, similar speeches, or multiplex connections.