A Methodology for Identifying Content Features in Popular Socio-Political Telegram Channels

This article presents a methodology for the automated analysis of content from popular socio-political Telegram channels and the results of its validation using empirical data from 2021–2024. The relevance of the study stems from the key role of digital platforms, particularly Telegram, in shaping the contemporary public agenda, as well as the need to develop reliable tools for studying mechanisms of network communication, such as social contagion in networks. The main goal was to develop and validate a comprehensive methodological approach capable of identifying stable content characteristics: thematic structure, keywords (including key named entities such as persons, organizations, locations, etc.), emotional and evaluative patterns (sentiment, emotions, degree of toxicity), and semantic connections between channels. Open, pre-trained models were used as the primary data processing tools. An assessment was made of the links between the named entities on the one hand and the type of sentiment and degree of toxicity on the other. In addition, graphs of thematic and contextual connections between the selected public Telegram channels were constructed. Despite a tendency toward the formation of distinct contextual clusters, popular socio-political channels are generally characterized by thematic homogeneity. Their content is presented predominantly in a neutral, unemotional tone, and topics that are clearly alarming for society dominate. The high degree of personalization of the events described is noteworthy: “opponents” are often designated not only as individuals but also as organizations. The results obtained can be further used to study the mechanisms of (1) the formation of new methods of political persuasion online, group polarization, and online activism under the influence of network social contagion; (2) thematic contagion; and (3) the formation, reproduction, and consolidation of cultural practices.