Extensive research has shown that unilateral optimization of transit systems is not effective enough to significantly increase its transport efficiency. Considering that urban land-use characteristics, including residential, work, consumption, transit, and so forth, are significantly interrelated with travel demands and travel behaviors, this paper provides a way to optimize transit system by raising awareness of the relation between ridership and built environment. This paper adopted point of interest (POI) data to investigate the effect of physical built environment on online car-hailing ridership in Chengdu, China. The study area was tessellated with several Voronoi cells; these cells were further clustered into three ridership patterns based on the time-varying characteristic of ridership. Given that some differences existed in the three ridership patterns, a separate spatial ridership model was developed to understand the factors that influence ridership patterns using geographic weighted regression (GWR) analysis. The data and results verified that the built environment had various influences on online car-hailing alighting ridership in spatial and temporal dimensions, of which the significant POI factors for determining the ridership pattern during different periods were detected. Remarkably, this study took the ridership dataset from the online car-hailing transit system, mainly because the pick-up (PU) and drop-off (DO) locations generated by this service are closest to the origin and destination of the trip, except that it is more popular recently. Therefore, the analysis of the impact of built environment on travel based on the online car-hailing dataset can be captured in greater detail, with a higher degree of accuracy.