Abstract

Social movements are dominated by storytelling, as narratives play a key role in how communities involved in these movements shape their identities. Thus, recognizing the accepted narratives of different communities is central to understanding social movements. In this context, journalists face the challenge of making sense of these emerging narratives in social media when they seek to report social protests. Thus, they would benefit from support tools that allow them to identify and explore such narratives. In this work, we propose a narrative extraction algorithm from social media that incorporates the concept of community acceptance. Using our method, we study the 2021 Cuban protests and characterize five relevant communities. The extracted narratives differ in both structure and content across communities. Our work has implications in the study of social movements, intelligence analysis, computational journalism, and misinformation research.