TRACE
Textual and Relational Analysis of Community Extremism in BLM Twitter Using BERT and GNNs
About TRACE
Online extremism is a growing challenge in the digital age, especially during times of political and social movements. This project introduces a hybrid BERT-GNN framework designed to detect extremist narratives and uncover relational patterns in Black Lives Matter discussions on X, formerly Twitter. Our system bridges text classification and network analysis to provide actionable insights for researchers, policymakers, and advocacy groups.
Researchers
Analyze extremist discourse trends with classification results and interaction graphs.
Policy Analysts
Identify extremist clusters, influential accounts, and risky narratives for policy insights.
Citizen / Journalists
Understand extremist influence patterns through clear, visualized networks.