Využití informací ze sociálních sítí pro podporu krizového řízení

Abstract

The thesis deals with the use of information from social networks to support crisis management, focusing on the analysis of data from the BlueSky platform. The theoretical part characterizes the crisis management system in the Czech Republic, describes the specifics of social networks as sources of information during emergencies, and analyzes the activities of virtual operation support teams (VOST). In the practical part, a solution in the Python programming language is designed and implemented, which extracts, processes, and analyzes posts from BlueSky using artificial intelligence technologies. The created tool uses pre-trained models for sentiment analysis, named entity recognition, and detection of potential disinformation. The obtained data are then stored in a PostgreSQL database and visualized through an interactive dashboard created using the Streamlit library. The work focuses on monitoring posts related to fires and evaluates their information value for crisis management. The results confirm the potential of social network analysis as an additional source of information for crisis managers, especially in the early detection of emergencies and monitoring public sentiment. Finally, the limitations of the proposed solution and possibilities for its improvement are discussed.