Spatial And Temporal Characteristics Analysis And Influencing Factors Of Shared Bicycle Riding

In recent years,urban bike sharing has developed rapidly,with the characteristics of green,low-carbon,low-cost,flexible and so on,which provides convenience for people to travel.However,while sharing bicycles improve transportation,beautify the environment and facilitate travel,there are also problems such as excessive delivery and disorderly parking.In this paper,SPSS statistical analysis and geographic information system(GIS)spatial statistical analysis of kernel density estimation,spatial autocorrelation,regression model,variance expansion coefficient,single sample K-S normal distribution test method.The research on the spatio-temporal characteristics and influencing factors of shared bicycle riding provides certain guidance for the deployment,distribution,scheduling,operation and management of shared bicycles in Mobike,Shanghai.The main research contents and results of this paper are as follows:(1).Spatial and temporal characteristics of shared bike ridingFirstly,this paper uses SPSS statistical analysis to analyze the riding time,riding time,turnover frequency and riding distance of shared bikes.The results show that in the time series,the overall trend of the daily dynamic characteristics of shared bicycle riding is M-shaped,and there are three peak periods of riding:early,middle and late.Riding time:76.58%of the users riding within20min,which riding6-10min users accounted for 31.58%;the average riding time of users is 16 min:Turnover frequency:9.26%of the bikes with turnover rate of 1-3 times a day,and 90.74%of the bikes with turnover rate of more than 3 times a day,which fully shows that Mobike is in good condition;Riding distance:less than3km is the main reference distance for residents to choose cycling,and2km is the optimal reference distance.With the help of ArcGIS spatial statistical analysis,the spatial distribution of bicycle riding,riding hotspots,bicycle riding sites and riding hotspots are deeply excavated.In terms of the current situation of bicycle riding,the number and activity density of bicycle riding in each unit are analyzed by spatial statistics.The results show that the number of bicycle riding is the largest in Yinhang Street,Wujiaochang Street and Dachang Town,and the activity density of bicycles is generally distributed in a circle,and gradually sparse from inside to outside.The global Moran’s I index,the generalized General G statistics and the local Moran’s I index are used to describe the spatial distribution characteristics of bicycle activities.It is found that the streets such as Daqiao Street,Dinghai Road Street,Yinhang Street,Wujiaochang Street,Xinjiangwancheng Street and Jiangwan Town show high aggregation characteristics,and the other regions show random distribution characteristics,indicating that the bicycle activities of most units are random.Based on the nuclear density estimation and Getis-Ord Gi*detection method,the hot spots of bicycle activities were extracted.It was found that the hot spots were continuous and regional,and mainly distributed in Yangpu District,Baoshan District,Hongkou District,Huangpu District and Putuo District.The number of cycling sites obtained by OD model and fishing net function is large and highly concentrated,which basically covers the main urban areas of Shanghai.There are 27846 rental return sites and 3885 cycling equilibrium sites,accounting for 13.94%of the total number of cycling sites,indicating that the scheduling optimization of bicycles is extremely important.Using the shortest path analysis function in the network analysis module(Network Analyst)and the network kernel density estimation,the hot sections of bicycle riding show the distribution characteristics of ’large aggregation and small dispersion’,and are concentrated in Yangpu District,Baoshan District,Hongkou District,and other areas.(2).Spatial correlation between shared bicycles and urban elementsFirstly,the data acquisition and preprocessing of the influencing factors that may affect the sharing bike riding activities are carried out.The data of roads,buildings,population,public transport and subway are selected as independent variables,and the number of bike riding in Mobike is used as the dependent variable.The variance expansion factor and Moran’s I index were used to test the independent variables and dependent variables.The least squares regression model and geographically weighted regression model were constructed for the selected independent variables and dependent variables.The fitting effects of the two models were compared and analyzed to explore the correlation between urban factors and bicycle activities.R2 after correction of the fitting index of the global regression model is 0.545;Model performance index AICc is 167.688;The Koenker(BP)value of the steady-state significance test of the model is 19.3985,which passes the significance test;The fitting degree and significance of the model show that there is spatial instability among variables,so the description of the model has certain limitations.Compared with the least square regression model,the R2 of the geographically weighted regression model is 0.810,and the interpretation ability is improved by 26.5%.Akaike Information Criterion(AICc)is 119.373,which is reduced by 48.315 compared with the OLS model,and the performance of this model is better.For the Residual Squares index that measures the fitting degree of sample data,the smaller the value is,the more the model fits the sample data.The Residual Squares values of OLS and GWR models are 31.832 and 11.004,respectively,indicating that the GWR model fits the sample data more.In summary,both the performance and fitting degree of the model prove that the GWR model is superior to the OLS model.The regression results of the model show that there is a certain spatial correlation between urban factors and bicycle activities,and this correlation has obvious spatial heterogeneity.The experimental results show that the average correlation degree of regression coefficient from high to low is construction,population,POI,subway,bus and road.Regression coefficient range from high to low is building,road,bus,railway station,POI,population.In the regression coefficient of the influencing factors,the standard deviation of the building is the largest.