Y. Deng, Y. Cao, X. Hu
In recent years, research on the relationship between the built environment and usage pattern of public bike-sharing has attracted some attention. However, studies that focused on the exploration of spatial-temporal heterogeneity of free-floating bike-sharing (FFBS) usage pattern were less common. To this end, based on the collection of GPS (Global Positioning System) data from the devices embedded on the Mobike shared bikes in Xi’an, China, this paper analyzed the spatial-temporal heterogeneity of built environment and FFBS usage pattern during the morning (7–9 am) and evening (5–7 pm) rush hours. The goals were to identify which variables would have an impact on FFBS trip due to different geographic locations, and analyze the impact of the same variable affected by geographic location on FFBS trip during morning and evening peaks. Geographically weighted regression (GWR) model was developed, and was found to achieve much better performance at the spatial location than the global models. GWR helped to better identify variables with temporal and spatial heterogeneity. The result showed that there was a high degree of spatial heterogeneity, which indicated the leading factors could be significantly different in various zones during rush hours. Cycling facilities were found to have a more significant impact on the trip generation in the morning rush hours than the evening rush hours, but had little impact on the trip attraction during the morning rush hours. In addition, when cycling facilities, land use and public transportation were used as explanatory variables, more evident relationships were observed between them and FFBS usage in evening rush hours than morning. In particular, the amount of transfer demand between bikes and subway was more obvious in evening rush hours. The results indicated that the policies and interventions should be more targeted to improve the rebalance of FFBS in the spatial-temporal geographical context.
Keywords: Free-floating bike-sharing (FFBS); spatial-temporal heterogeneity; geographically weighted regression (GWR); built environment; GPS (Global Positioning System) data; China