LDA 토픽모델링 기반 뉴스 분석을 통한 청도군 지역축제 이슈와 함의

초록

This study analyzes 35 years of news coverage (since 1990) on local festivals in Cheongdo-gun, Gyeongsangbuk-do, using Latent Dirichlet Allocation (LDA) topic modeling to identify key thematic trends and derive policy implications. Seven major topics emerged, including traditional cultural events, cultural and artistic activities, community development, critical issues such as animal cruelty, and promotion by intermediary organizations and so on. Visualization through the pyLDAvis revealed tensions between cultural symbolism and policy implementation, as well as varying degrees of public acceptance and conflict. The findings The findings could emphasize the need to move beyond flagship festival-centered strategies toward diversified content and more horizontal governance. Policy suggestions include enhancing the brand value of the Cheongdo Bullfighting and Persimmon Festivals, strengthening the roles of intermediary institutions, and promoting eco-friendly festivals in connection with the Saemaul Undong. By adopting a big data-driven approach, this study provides a structural understanding of local festivals and addresses the limitations of traditional case study and survey-based methods.

키워드