The Influence of Community Interactions on User Affinity in Social Networks: A Facebook Case Study

With the advent of social media, one of the biggest concerns has been its suspected impact on democracy by influencing users' opinions during elections. The accusations vary from the propagation of fake news to concerns about Facebook having unfettered power over its users' content. As a first step towards understanding the true nature of social media's influence, we started collecting data prior to Local Election in Sri Lanka, 2018 and continued the collection until the election was over; we collected the data through Facebook API and all of the data is completely anonymized. The dataset covers 44K users from Sri Lanka and their interactions with 44 Facebook groups. The dataset also includes all of the associated events related to each group. As a preliminary step, we have analyzed how the user affinity surrounding these groups have changed surrounding the period of the local election. Our analysis also gives a concrete and quantitative evidence of how users are communicating on Facebook and how active they are surrounding a sensitive event. Further, there is a significant change of affinity of a set of individuals over the time corresponding to the external event. We believe our analysis is the first of its kind in providing scientific analysis of the influence of social media in Sri Lanka.