Analysis Of Bike-sharing Travel Characteristics Under Various Season Conditions

Bike-sharing is a green,healthy and convenient way of travel mode.As a supplement to urban green public transport,it has many advantages in aleviating traffic congestion and facilitating the passengers.In this paper,Yixing City is selected as a typical representative of small and medium-sized cities.By acquiring and integrating multi-source data,the differences of bike-sharing travel characteristics and their influencing factors in different seasons are studied.Firstly,the bike-sharing smart card data and weather data for a whole year of Yixing city are acquired,and longitude and latitude information of bike-sharing stations and road information data are obtained from the internet.Then,the invalid data,redundant data,and system error data in the original data were discriminated and rejected.Subsequently,the GIS data are imported into the GIS software to obtain the POI number and road feature information in the 300-meter buffer zone of the bike-sharing stations.Further data processing and data fusion are carried out by using Python language,which provides the analysis basis for the follow-up study.Secondly,the multi-source fusion data are visualized by Python,Microsoft Excel and ArcGIS software.Combined with the visualization results,the characteristics of travel time distribution,travel distance distribution,user travel frequency and frequency characteristics,spatial distribution characteristics,and different gender and age groups of users are discussed.Furthermore,with each bike-sharing station as the regression point and the use of usage volume at each bike-sharing station as the dependent variable,the traditional linear regression model(OLS)and geographically weighted regression model(GWR)are constructed to explore the relationship between usage of bike-sharing and social demographic and land use factors on weekday/weekend and morning/evening peaks under various weather conditions.From the fitting results of the model,GWR is significantly better than OLS.Taking variables residential POI and bus station density as the examples,the estimation results of parameters are visualized and analyzed.Finally,based on the research results and the bike-sharing travel characteristics under various season conditions in Yixing city,some policy suggestions are put forward to improve the attractiveness of bike-sharing from the aspects of improving bike performance,bike station layout and bike dispatching,respectively.