Discovering Structures of Communities in the #StopHateForProfit Network: A Social Network Analysis

Abstract

The boycott campaign against Facebook #StopHate - ForProfit, launched in June 2020, emerged as a key phe - nomenon in the fight against hate speech on social media. This study addresses the detection and characterization of communities in the #StopHateForProfit campaign, employing theoretical and methodological approaches from Social Network Analysis ( SNA ) and Natural Lan - guage Processing ( NLP ) to examine the social structure of the campaign on Twitter (now X). We used the software Gephi for community detection, employing centrality, modularity, connected components, and clustering coef - ficient measures. The analysis disclosed a complex and cohesive network composed of 5,556 communities with a high modularity that indicated dense internal interac - tions. We identified the strongest and weakest connected actors in the communities, which hinted at the closest and most direct relationships. The classification of actors ac - cording to their position provided insight into node influ - ence and cohesion in the network. This interdisciplinary line of action contributes to understanding the diversity of approaches within the #StopHateForProfit campaign, highlighting its relevance regarding mass participation and impact. The analysis of communities revealed an ef - fective collaboration among actors, demonstrating the comprehensiveness of the coordinated strategy to counter hate speech