Algorithmic Analysis & Visualization Of Bots’ Influence On US Elections & Your Vote

Abstract

Through analyzing large corpora of Twitter data collected during the 2016 US Primaries, we map the online debate surrounding elections via a force-directed algorithm - creating novel visualizations of the election-related debate. Moreover, using machine learning & statistical approaches, we identify bot accounts and assess their characteristics and impact on sentiment. Concerning bot characteristics, we find a correlation between a user’s avg. sentiment and botlikeness and show that bots primarily target influencers with 300k-1M followers. Moreover, we find that bots impact sentiment as effectively as humans over single interaction, contribute to sentiment silos, and preserve hostility more than their human counterparts when sharing content.