Dynamics of Russian anti-war discourse on X (Twitter): a computational analysis using NLP and network methods

Abstract This paper explores the dissemination and impact of Russian anti-war discourse on X (Twitter) following Russia’s 2022 invasion of Ukraine. It examines the dynamics of this discourse within the Russian-speaking segment, identifies key actors and communities, assesses the role of bot accounts, and analyzes dominant narratives and communication strategies through framing analysis. The methodology employs a mixed-methods approach, including data collection via the Twitter API, social network analysis, bot detection, community detection, natural language processing with BERTopic, and framing analysis using the BEND model. This includes examining persuasion tactics in two distinct discourses: one involving anti-war opposition accounts and the other centered on pro-government narrative accounts. The findings provide insights into the dynamics of Russian anti-war discourse on X (Twitter), highlighting prevalent narratives and communication strategies.