Exploring Influencing Factors On Stop-Level Transit Ridership Considering Endogeneity And Spatial Heterogeneity

In recent years,the bus ridership in many cities has generally declined,and it is of great practical significance to explore the influencing factors of bus ridership in order to improve the ridership.The bus ridership is affected by many internal and external factors.The traditional multiple regression model is often used in the analysis of influencing factors of bus ridership because of its simple application and easy explanation.However,when explanatory variables in the model are correlated with error terms,endogeneity will occur,making the model estimation results biased and inconsistent.As a global model,the multiple regression model does not consider the difference of the influence degree of the same variable on different regions,that is,the influence of spatial heterogeneity is ignored.This paper studies the bus ridership at the stop level.Based on the multiple regression model,considering the endogeneity from the overall perspective and the spatial heterogeneity from the local perspective to analyze the impact of various factors on the ridership,so as to provide reference for effective policy-making and management decision-making.This paper takes the six districts of the main city and Jiangbei New District of Nanjing as examples.After sorting out the factors affecting the bus ridership at the stop level,this paper obtains and processes multi-source data such as points of interest,residential areas,road network,land use,and public transportation,and establishes a data set of influencing factors including land use,transportation infrastructure,station attributes and socio-economic factors.The multiple regression models of the morning peak period,evening peak period and whole day period on weekdays were established,and the significance and importance of the variables were briefly analyzed.Considering the endogeneity caused by the bidirectional causality between the headway and the bus stop ridership,a simultaneous equation model was established.After analyzing and summarizing the results of parameter estimation,suggestions for improving ridership were put forward.Based on exploratory spatial data analysis and considering the influence of spatial heterogeneity,a geographically weighted regression model was established,and the parameter estimation results were analyzed visually.The results show that: 1)the multiple regression model without considering endogeneity underestimates the impact of headway on ridership in each period.2)From the overall level,the headway and the number of lines are the most important factors affecting the bus stop ridership.Commuters have a great impact on the bus ridership.Public bicycle stops have a positive impact on the bus ridership in the evening peak period of weekdays.3)From a local perspective,the same variable has different degrees of influence in different spatial locations.The influence degree of headway decreases outward with Xuanwu Lake area in Nan Jing as the center.The number of station lines has the least influence on the old urban area and the greatest influence on the peripheral areas.Differentiated ridership promotion strategies can be implemented for different regions.The research results of this paper can provide a theoretical basis and reference for public transport related planning,management and policy formulation,and have a certain significance to improve the attractiveness of public transportation.