Monitoring public sentiment is essential for government authorities at all levels. Although various qualitative and quantitative research methods have been developed to study citizens’ electoral behavior, these methods often lack the granularity needed to pinpoint region-specific problems, especially in large federations such as the Russian Federation. This paper proposes a two-stage framework that combines a pattern clustering of electoral behavior data from the 2023 gubernatorial elections in Russia with an analysis of social media posts. First, we group the voting precincts (UIK) into patterns using pattern clustering aidentify theify the regions showing protest activity, defined by the proportion of votes for the winning party falling below a certain threshold. Second, we apply hierarchy clustering (based on cosine distance) of the comments on social networks (VKontakte, Telegram, Odnoklassniki, Twitter) to uncover specific socioeconomic issues. Four regions—Voronezh Oblast, Krasnoyarsk Krai, Altai Krai, and Primorsky Krai—exhibited patterns of strong opposition or protest. Further text analysis of their social media data indicates overlapping problems, including deficiencies in healthcare, water supply, and delayed wages. We discuss how combining pattern electoral clustering with social media text mining can be integrated into ongoing public opinion monitoring systems to better detect and address emerging regional problems.