Alexandre Bovet presented a talk titled "Twitter Influencers and Increased Media Polarization during two U.S. Presidential Elections," co-authored by James Flamino, Brendan Cross, Boleslaw K. Szymanski from NEST at the Networks 2021 Satellite

Alexandre Bovet presented a talk titled "Twitter Influencers and Increased Media Polarization during two U.S. Presidential Elections," co-authored by Stuart Feldman, James Flamino, Alessandro Galeazzi, Brendan Cross, Zhenkun Zhou, Matteo Serafino, Hernán A. Makse, and Boleslaw K. Szymanski at the Political Communication Networks PCNet21 at Networks 2021 Satellite Workshop on June 24, 2021. Traditionally, political party elites define the political program and distribute it to the supporters through the media and other communication channels. The decentralized and fast reacting to events social media freely offering users both consumer and producer roles have changed these dynamics. Embraced by traditional elites, social media also enabled an emergence of new influencers outside of such elites. To study the effects of this change, we analyze Twitter interactions during the 2016 and 2020 US elections. We find that new entrants have a strong impact, comparable to more famous or established individuals and organizations. We observe that numerous of their sites gained strong followings indicating their significant influence on the public. Yet, many of them were distributing low quality information as measured by a number of metrics. We apply a number of computational techniques to extract information from massive amounts of interactions on Twitter and measure polarization of these diverse influencers during the studied US presidential elections. In particular, we investigate the evolution of the news diffusion and the change in news influencers. by classifying 2020 tweets with a link to a news outlet in several categories, reconstruct the networks of news diffusion and identify the news influencers. We compare these results with the ones obtained in 2016 and then investigate the evolution of the polarization between news influencers and between Twitter users from 2016 to 2020. We identify which influencers are linked to media organizations and political parties, and which accounts are entirely independent. We infer the ideology of Twitter users based on the political actors they follow by projecting the bipartite network between influencers and all the Twitter users that retweeted them on a latent ideology scale. The ideology distribution reveals an increased polarization of the influencers and their followers from 2016 to 2020 (measured by Hartigan's dip test). Our results reveal the emergence of a new type of social media influencers, different from the traditional political elites, accompanied with an increased polarization of the Twitter users.