This repo supports the paper "Leveraging Secure Social Media Crowdsourcing for Gathering Firsthand Accounts in Conflict Zones." We collected 250,000 Telegram messages on the Russo-Ukrainian conflict and developed a framework to identify contextual groupings, revealing 477 new insights from 13 news sources.
nlp natural-language-processing social-media osint telegram text-analysis data-collection crowdsourcing computational-social-science conflict-analysis social-media-analysis asonam-2024
-
Updated
Sep 1, 2025 - Python