Abstract (eng)
This thesis investigates the utilisation and ridership patterns of public transport in Bratislava, with a focus on the relationship between weather conditions and passenger numbers across buses, trolleybuses, and trams. Its methodology is grounded in the CRISP-DM model. The modelling involved training SARIMAX for time series and XGBoost for predictive analysis. The results reveal that temperature has a modest impact on bus ridership on weekends, leading to a slight increase in number of passengers, while precipitation has a significant negative impact on ridership of both buses and trolleybuses. Humidity minimally affects bus usage but significantly impacts trolleybuses on weekdays. Tram ridership declines during weekends with higher temperatures and precipitation. School and public holidays also exhibit a notable decrease in ridership across all modes of transport. Lastly, the study offers actionable strategies for urban transportation planning, emphasising data-driven decision-making such as employing resource allocation based on temporal and weather-related patterns to anticipate ridership shifts. In summary, this research contributes to understanding the impact of weather conditions on public transport ridership, providing valuable insights for the enhancement of Bratislava's transit system.