Extracting significant signal of news consumption from social networks: the case of Twitter in Italian political elections

Abstract: According to the Eurobarometer report about EU media use of May 2018, the number of European citizens who consult on-line social networks for accessing information is considerably increasing. In this work we analyze approximately $10^6$ tweets exchanged during the last Italian elections. By using an entropy-based null model discounting the activity of the users, we first identify potential political alliances within the group of verified accounts: if two verified users are retweeted more than expected by the non-verified ones, they are likely to be related. Then, we derive the users' affiliation to a coalition measuring the polarization of unverified accounts. Finally, we study the bipartite directed representation of the tweets and retweets network, in which tweets and users are collected on the two layers. Users with the highest out-degree identify the most popular ones, whereas highest out-degree posts are the most "viral". We identify significant content spreaders by statistically validating the connections that cannot be explained by users' tweeting activity and posts' virality by using an entropy-based null model as benchmark. The analysis of the directed network of validated retweets reveals signals of the alliances formed after the elections, highlighting commonalities of interests before the event of the national elections.