Digital Traces of Extremist Narrative Construction

I’ll be presenting some ongoing work analyzing digital trace data to characterize how online extremist spaces build narratives. I combine temporal social network analysis with temporal topic modeling to reveal how narratives spread and amplify over these networks. In the cases studied, narratives amplified to a point where they became topics of mainstream discussion. I look specifically for patterns preceding the break into mainstream. I’ll situate these case studies within the broader goal of my dissertation, which is to describe, and to some extent, explain the online ecosystem of misinformation and extremism. I frame the way its entities interact with one another in a technologically-mediated environment as a complex adaptive social system, which I use NLP, network science, and other relevant tools to interrogate.

Content warning for the audience — I specifically look at how these spaces target vulnerable groups like immigrant communities and LGBTQ folks. I’ll only show aggregate analysis, no full posts or even sentences. I’ll also censor offensive phrases, but I still completely understand anyone preferring to sit this one out.