Detecting Political Polarization Using Social Media Data

Political polarization is a world wide problem observed everywhere. It is on the increase due to increasing digital communication and use of social media. Election times are especially important, increasing political polarization tremendously, causing social and political problems in the society. In this study, we take the case of last Turkish local elections, which was highly contentious, leading to the renewal of elections in the most-populated major metropolitan city of Istanbul with a population of 15 million people. We used Twitter data to measure the political polarization, collected over 90 million tweets and retweets, introduced the inter retweet ratio metric to quantitatively measure the political polarization. Our results show that political polarization increases tremendously by double digit decreases in inter community interactions towards the election days, and comes back to normal after the elections are over.