Spatiotemporal Patterns and Influencing Factors of Shared Bikes Use for Metro Access and Egress Trips: a Case Study in Xi’An, China
Junqi Sun
2020-01-01
Abstract:Dockless bike-sharing system (DLBS) as a potential mode to solve the “first and last mile” problem is commonly used as access and egress of metro systems in many Chinese cities. A better understanding of this novel, economical and environmentally friendly mode of public transport can help improve the efficiency of bike-sharing system, and enhance the connection between bike sharing and metro systems so as to bring convenience to the travellers. To do so, this study used different types of visualization techniques, based on the large-scale shared bike data, to explore the spatial and temporal use patterns of shared bikes use for metro system in Xi’an city, China. Besides, the geographically weighted regression (GWR) model was applied to reveal the relationship between the influencing landuse related factors and the use patterns of shared bikes. Specifically, the Python-based data process method was used to deal with the large-scale GPS data of the shared bikes to extract useful information, such as the origins and destinations of the bike-metro trips. The cycling trajectory was also generated based on “Rout Plan” Application programming interface (API) from Baidu Map Server. A variety of statistical chart combinations, as well as different visualization methods such as heat maps, interactive flow maps, and routes distribution maps were used to extract the hot spots of the shared bike use. With the help of different visualization software or platforms, the spatial characteristics of the shared bike use in morning and evening peak hours were identified. In short, in the morning rush hours, there are a large number of shared bike trips that ride from the residential area to the metro station, or from the metro station to the job locations. While, in the evening peak hours, a large number of shared bike trips depart from the employment concentration to the metro station, or ride from the metro station to the residential area. This trend is consistent with the commuting activities on weekdays. Point of interest (POI) data was used as the basis for the influencing land use related factors The number of residence, job, commercial area, recreation area, green space, educational place, and health care locations were used as independent variables to explore their influence on the number of origins and destinations of bike-metro trips during different time slots. From the geographically weighted regression (GWR) model, in the morning peak, the number of residences has a positive effect on the distribution of the origins of the bike-metro trips, and the impact has spatial heterogeneity. The number of job locations and educational places has a positive influence on the distribution of destinations of bike-metro trips, and the influence has spatial heterogeneity. In the evening peak hours, the number of job locations has a positive impact on the distribution of the origins of bike-metro trips, and the influence has spatial heterogeneity. The number of recreation areas and residences has a positive effect on the distribution of the destinations of the shared bike use with spatial heterogeneity. According to the results of this research, some suggestions were provided to improve the efficiency of the bike-sharing system so as to make shared bikes solve the first mile/last mile problem better.