Nonlinearities and threshold points in the effect of contextual features on the spatial and temporal variability of bus use in Beijing using explainable machine learning: Predictable or uncertain trips?

In pursuing sustainable transport, understanding the dynamics of transit passengers' travel demand is necessary for establishing more attractive public transport relative to cars. However, to what extent daily transit use displays geographic and temporal variabilities or predictability, and identifying what are the contributing factors explaining these patterns have not been fully addressed. Drawing on smart card data in Beijing, China, this study adopts new indices to capture the spatial and temporal variability of bus use during peak hours and investigates their associations with relevant contextual features. Using explainable machine learning, our findings reveal non-linearities and threshold points in the spatial and temporal variability of bus trips as a function of trip frequency. Greater distance to the urban centres (>10 km) is associated with increased spatial variability of bus use, while greater separation of trip origins and destinations from the subcentres reduces both spatial and temporal variability reflecting highly predictable of trips. Higher availability of bus routes is linked to higher spatial variability but lower temporal variability. Meanwhile, both lower and higher road density is associated with higher spatial variability of bus use especially in morning times. These findings indicate that different built environment features moderate the flexibility of choosing travel time and locations influencing the predictability of trips. Understanding highly predictable trips is key to develop more effective planning and operation of public transport.