Title:Mining Influentials and their Bot Activities on Twitter Campaigns

Abstract:Twitter is increasingly used for political, advertising and marketing campaigns, where the main aim is to influence users to support specific causes, individuals or groups. We propose a novel methodology for mining and analyzing Twitter campaigns, which includes: (i) collecting tweets and detecting topics relating to a campaign; (ii) mining important campaign topics using scientometrics measures; (iii) modelling user interests using hashtags and topical entropy; (iv) identifying influential users using an adapted PageRank score; and (v) various metrics and visualization techniques for identifying bot-like activities. While this methodology is generalizable to multiple campaign types, we demonstrate its effectiveness on the 2017 German federal election.