Digital preservation of academic heritage: a computational analysis of the Agora Discussion Club at Kryvyi Rih State Pedagogical University (2016-2025)

Abstract Regional academic discussion clubs represent valuable but understudied resources for understanding intellectual discourse in higher education. This study presents a comprehensive computational analysis of the Agora Discussion Club, an academic forum operating within the History Department at Kryvyi Rih State Pedagogical University, Ukraine, from October 2016 to January 2025. Using web scraping, optical character recognition (OCR), machine translation, and natural language processing techniques, we constructed a bilingual corpus of 80 discussion sessions spanning nearly a decade of academic activity. Our enhanced named entity recognition pipeline, combining spaCy's Ukrainian model with morphological normalization and OCR-enriched session data, extracted 714 unique entities across five categories (445 persons, 134 locations, 121 organizations, 7 events, 7 works). Network analysis identified 32 recurring presenters, with the most active moderator appearing in 14 sessions. We tracked thematic evolution across eight academic years, including the significant disruption of Russia's full-scale invasion in February 2022. The corpus reveals diverse topical coverage from ancient history to artificial intelligence, with engagement patterns showing correlations between military and technology themes and higher viewership. This work demonstrates the feasibility of applying digital humanities methods to preserve and analyse regional academic heritage, particularly from institutions operating under conditions of conflict. The complete dataset is released under CC BY 4.0, providing a foundation for future research on Ukrainian academic discourse and digital preservation methodologies.