Spatio-Temporal Investigation of Public Transport Demand Using Smart Card Data

Abstract Policymakers must find efficient public transport solutions to promote sustainability and provide urban mobility in the course of growth. A growing number research papers are applying Geographically weighted regression (GWR) model relationship between demand its influential factors. However, few studies have considered rapid development journey inference from ticket transaction data. Similarly, potential GWR analyze spatio-temporal changes that reflect transportation supply thus a measure for evaluating local success has yet be exploited. In this paper, we use inferred journeys smart card inferences as dependent variable how responds set explanatory variables, emphasizing supply. Consequently, successor Multiscale Weighted Regression (MGWR) applied spatially varying impact seven consecutive time frames autumn 2017 spring 2020, allowing conclusions about demand, well benchmarking changes. The (M)GWR framework’s predictive power is evaluated by training with past data testing following years. conducted analyses reveal model, using data, can retrospectively predict on travel behavior provides policies.