Leave or Remain? Deciphering Brexit Deliberations on Twitter

Apalak Khatua, Aparup Khatua

Abstract: Twitter is becoming a popular microblogging platform for exploring various socio-political movements. We extract 2.7 million Brexit related tweets to decipher the European Union (EU) referendum deliberations. Our volumetric analysis correctly predicts the outcome of 2016 Brexit referendum. We also investigate whether Twitter discussion adequately reflects the socio-economic concerns related to the UK's decision to leave or remain in the EU. We consider hierarchical clustering analysis (HCA) to explore various underlying themes of this political discourse. To tackle a large portion of retweets in our corpus, we employ tanglegram framework to graphically compare HCA of unique tweets (after removing the duplicates tweets) and the entire corpus. We note finer variances between these two HCA.