| With the characteristics of low-carbon,environmental protection and flexibility,bike sharing is an important way to solve the "last mile" of subway travel,effectively expanding the service range of subway stations and increasing the accessibility of the subway.In-depth analysis of the travel behavior of bike sharing connecting with metro stations will help to improve the efficiency of bike sharing to metro,improve the accessibility of the subway,and realize efficient and sustainable urban transportation.Therefore,taking Nanjing as a case study,based on the multi-source data such as bike sharing order data,metro line and station data,road network data and POI data,the author identifies the travel records of connecting subway,deeply analyzes the travel characteristics of bike sharing connecting with subways,and explores the impact of built environmental factors on the connecting travel volume of different subway stations.Firstly,the acquisition method and pre-processing process of multi-source data required by the research are introduced in detail,so as to provide data basis for subsequent analysis.Subsequently,the travel behavior of bike sharing in different areas around the subway station is deeply analyzed,and a variety of bike sharing travel types are classified.Then,the method of identifying the connecting travel is proposed.The identification methods are as follows:first,a multi ring buffer is constructed for each entrance and exit of the subway station,and the optimal buffer radius is determined according to the changes of the increase of bike sharing orders within the multi ring buffer.Second,based on the coordinates of the starting and ending points of bike sharing,the possible feeder trips within the buffer zone are extracted.Third,some non-connected trips are excluded based on the spatial location of the starting and ending points,subway lines and riding distance constraints,and the rest are connecting trips.Based on the bike sharing order data,metro line and station data in Nanjing,the identification method proposed in this paper is used to extract the connecting trips in Nanjing.Compared with the traditional buffer zone identification method,the identification method proposed in this paper eliminates about 25% of the non-connected data and significantly improves the accuracy of connecting trips.Secondly,based on the identification results,the connecting trips are divided into four categories according to the connecting time and connecting type: weekday-access,weekday-egress,weekend-access,and weekend-egress.The temporal and spatial distribution patterns of different types of connection trips are analyzed from the perspectives of travel volume,utilization rate of bike sharing,connecting distance and connecting duration.The comparative analysis shows that there are obvious differences in the volume of connecting trips at different periods.The connecting trips on weekdays show a significant peak in the morning and evening,while non weekdays are relatively stable.The supply and demand of connecting trips at different metro stations are unbalanced.The number of connecting trips is large and the utilization rate of bikes is low at downtown metro stations,while the number of connecting trips is low and the utilization rate is large at suburban metro stations.Finally,the global regression model(OLS),geographically weighted regression model(GWR)and multi-scale geographically weighted regression model(MGWR)are constructed to reveal the impact mechanism of connecting trips from the aspects of socio-economic attributes,land use,road infrastructure,bus accessibility and metro station attributes.The results show that the MGWR model,which considers the spatial heterogeneity and spatial scale difference of various factors,has a better fitting effect and explanation ability,and can better reveal the influence of various factors on connecting trips in different regions.In terms of land use,the number of business buildings and residential quarters has a significant positive impact on weekdays,and the impact of business buildings shows spatial non-stationary.The number of sports leisure,shopping and schools is positively correlated with connecting trips on weekend and shows strong spatial heterogeneity.In addition,combined with the research conclusions,the optimization strategies and suggestions of the bike sharing connecting with metro are put forward. |
