Identificazione di Echo-Chambers e comunità Gateway su Reddit tramite un approccio Graph-based

In recent years, online platforms have attracted large and vibrant commu-nities that changed how people interact and communicate. Platforms likeReddit provide a safe harbour to large communities to discuss a myriadof topics. Entire political movements and conspiracy theories started andspread on Reddit, leveraging controversy and polarization. In this thesis weinvestigate the dynamics and the flow of users across Reddit communities,called subreddits. The proposed methodology uses graph theory to studyconnections among subreddits, based on users temporal dynamics. In orderto analyze such a vast amount of data, in fact, we developed a graph-basedapproach, that efficiently summarize how Reddit users flow from one sub-reddit to the other over time. We found that our method captures emergingcommunity structures on Reddit: through ad-hoc hierarchical clusterings,we identify different polarized and temporarily stable clusters within politi-cized subreddits. Such clusters capture aligned communities, as validatedby a human-curated dataset. We employ this concise representation in or-der to analyze in detail border communities, i.e. gateway between differentpolarized clusters. We further characterize border communities as symmet-ric or asymmetric, identifying their different roles within the network. Theconstructed network represents, in summary, a comprehensive frameworkto study the dynamical behavior of users on Reddit. We show how it canhelp to identify how conspiracy groups grow and attract users, providing afoundation to design mitigation strategies to prevent the spread of harmful conspiracy theories.