Geoparsing vybraných sociálních sítí k veřejné dopravě

Abstract

This bachelor’s thesis focuses on the localization of tweets without embedded geolocation and their subsequent analysis in relation to public transport in selected cities of the Czech Republic. The main objective was to design and test a tool capable of automatically recognizing the names of transport stops within textual content and assigning them corresponding geographic coordinates. The work combines a rule-based approach to geoparsing with the use of official transport data (GTFS). The results were further visualized using maps that illustrate the spatial distribution of social media users‘ sentiment. The analysis confirmed the existence of recurring spatial patterns of sentiment. Negative sentiment tended to dominate in city centers and at transfer hubs, while more positive reactions were more common in peripheral areas. This work provides a new perspective on the use of social media data for evaluating the quality of urban transport and highlights the potential of text analysis for spatial research.