Research On Influencing Factors Of Transit Ridership Based On Multi-source Geographic Big Data In Guangzhou

With the development of economic and urbanization,urban population and car ownership is increasing every year.Traffic jams are a growing problem in many cities.Urban rail transit has become the most effective way to alleviate urban congestion due to its characteristics of fast speed,large volume,high frequency,punctuality and high comfort.At present,China’s rail transit planning and construction has entered a stage of rapid development.The planning scale continues to grow and gradually enters the network stage.Studying the influencing factors of rail transit ridership,it’s helpful for the planning for new transit stations,the management of established station ridership,the coordinated development of rail transit and land use.Based on multi-source geographic big data(including WeChat user data,building footprint and storey data,high-resolution images data and POI data),using backward stepwise regression,geographically weighted regression and K-means clustering to explore the influencing factors and the spatial differences of multi-type station ridership in Guangzhou city.(1)The fine built environment around the transit station can be identified by the fusion of multi-source geographical data.The relationship between population density,employment density,fine built-environment factors,fine land use,station content and multi-type station ridership.(2)It was found that four factors,including population density,floor area ratio,number of station entrances/exits and transfer dummy,had a significant positive impact on station ridership.There are obvious difference between weekday and weekend ridership model,between Morning-in & evening-out and Morning-out & evening-in,between Morning-peak and Evening-peak ridership model.(3)The spatial heterogeneity of station ridership influencing factors was investigate based on geographically weighted regression methods.The obvious difference influence was found between population density,employment density and traffic station site on the multi-type transit ridership.(4)The spatial difference of ridership influencing factor was explored based on K-means algorithm.At the end,the suggestions of different cluster zones were come up.