| With the vigorous development of the Internet economy,ridesourcing has become an important part of the modern public transportation system.The time from the passenger sending the request to the driver arriving at the passenger pick-up point is defined as the waiting time.As one of the standards of service level,it can reflect the accessibility of vehicle service.Therefore,studying the spatial distribution of waiting time and its influencing factors is of great significance in determining the location of low accessibility and highlighting traffic inequity.This can help policy makers and online car-hailing operators to improve ridesourcing service and accessibility.This paper analyzes the relationship between the waiting time for ridesourcing in the city,the social population economy and the built environment,and studies its changes in different units by using the data of ridesourcing orders in Austin combined with the local social and economic data and the built environment data.Considering spatial heterogeneity and spatio-temporal heterogeneity,respectively,a geographically weighted regression model(GWR)and a spatiotemporal geographically weighted regression model(GTWR)of waiting time for ridesourcing are established,and the spatial changes of estimated parameters are visualized.Specifically,firstly,use Python and GIS software to modify,filter and calculate the base map and basic data of the research area;secondly,build a geographically weighted regression model to explore the spatial heterogeneity of online ride-hailing waiting time,and estimate the parameters of the model Visualize spatial changes and analyze in-depth the impact of various influencing factors on the waiting time of ridesourcing in local areas.Finally,in view of the temporal and spatial heterogeneity of order data generated at different times and places,a temporal and spatial geographic weighted regression model is proposed to analyze the temporal and spatial changes of different influencing factors.The research results show that: 1)The fitting result of the spatial-temporal geographic weighted regression model is better than that of the Global regression model and the Geographic weighted regression model;2)The waiting time has an obvious time distribution law,the morning and evening peak waiting time is higher than the average peak,the night waiting time is lower than the day time;3)The proportion of ethnic minorities and the average income are related to morning and evening Peak waiting time has no effect,indicating that Austin local ridesourcing services are not limited to white and wealthy areas;4)The family car-free rate has a negative impact on the waiting time in the morning peak on working days,but has no effect on the evening peak;5)In areas with high road density and close to the city center,the waiting time is shorter,and it is easier for ridesourcing to enter these areas.This fact proves the importance of proper urban structure for the accessibility and efficiency of transportation services;6)As the results indicate that urban land use intensity has no significant impact on ridesourcing accessibility,a sustainable and less intense urban land use pattern(e.g.,more green spaces)may not hamper the fruit of the‘sharing economy’;7)Areas with a high density of public transit stations have shorter waiting times,and the areas covered by ridesourcing and public transportation services are the same in the city center.This makes it easy for people to leave the city center and difficult to enter the city center.For this observation Enlightenment for policy makers and regulators,strengthening the cooperation between ridesourcing and public transportation can provide a more balanced network between the entire city. |
