Recent events have led to a burgeoning awareness on the misuse of social media sites to affect political events, with the goal of
swaying the public opinion and confusing the voters. It is critical to
be able to comprehend the dynamics within these sites and spot users
polluting the information space. This has motivated a large body of
works on bots and troll detection, fake news detection, etc., which
mostly focus on classifying at the user level based on the content and
activities generated by the users. In this study, we take a higher-level
approach and propose SPG to map out all the activities in these social
media platforms around a broad topic and identify groups of users
that exhibit coordinated activity in polluting the general discourse around the topic of interest. To show the effectiveness of SPG, we
employ SPG to summarize and explain the polluting activity in
Twitter around the 2019 Canadian Federal Elections, by analyzing over 60 thousand user accounts connected through 3.4
million connections and 1.3 million hashtags. We verify that users in
polluting groups detected by SPG-flag are over 4x more likely to
become suspended while a majority skip the filtering in place. We
also verify that polluting hashtags discovered by SPG-signature
are known to be linked to misinformation campaigns. Finally, we
also show that a large block of right-wing conservatives based in
the US is heavily engaged in Canadian politics.

Organisé par la Chaire de recherche en études électorales et la Chaire de recherche du Canada en démocratie électorale.

DATE : Mardi 4 février, 12h-13h

LIEU : Salle C-4145, Pavillon Lionel-Groulx