Measuring political polarization using big data: The case of Turkish elections

Big data has been the driving force behind the latest machine learning and deep learning accomplishments in many learning tasks. Social media data, as a big data resource, has recently been used in many social studies to understand the social movements and political and social changes. In this study, we will analyze social media (Twitter) data to measure political polarization, which is one of the recent concerns in politics. This study made use of Twitter data collected in the 2019 elections in Turkey; new metrics are developed to measure the political polarization. We analyzed the political groups in the social network and then measure political polarization overtime during the election period. By applying community detection algorithms, we first identify communities based on the interactions among users. Then, we measure the interaction among user groups (communities) to successfully show the existence and growth of political polarization using big data during a general election process. To the best of our knowledge, this is the first wide-scale big data study on political polarization in a political election process.