Analysis Of Accessibility Of Urban Public Transport System Based On Multi-source Big Data

With the acceleration of China’s urbanization process,limited road resources and increasing traffic demand have gradually become important problems faced by the development of transportation system at this stage.In order to alleviate road pressure and imbalance between supply and demand and promote sustainable urban development,the state has vigorously supported the development of urban public transport system in terms of policies and funds to improve the travel rate of public transport.Accessibility of public transport system is an important index to measure the service level of public transport and guide the perfect management of public transport system.From the perspective of multi-source big data,this study models and evaluates the accessibility of urban public transport system.Firstly,based on many accessibility studies at home and abroad,this paper discusses the definition,characteristics and common evaluation models of accessibility.Based on the combination of multi-source big data,this paper puts forward the dynamic evaluation model of accessibility of urban public transport system and the three-stage prediction model of personal accessibility of public transport system.The main work and innovations of this paper are summarized as follows:(1)Based on traffic multi-source big data,the meaning of variables in the traditional gravity model is replaced,and an improved accessibility dynamic evaluation model is proposed.The dynamic information extracted based on multi-source big data is used to replace the static information of the traditional gravity model,so that the variables contained in the improved gravity model have space-time characteristics.(2)Taking Shenzhen,China as the research area,the urban area is divided into grids to simplify the analysis process,and analyze the time-varying law and spatial distribution characteristics of dynamic and static indicators and accessibility of public transport system.The conclusion shows that the allocation of public transport resources is guided by traffic demand,so the bus scheduling density and travel demand show similar time-varying characteristics.Regional land use has an essential impact on bus accessibility.The higher the degree of land development,the more bus facilities,and the higher the bus accessibility.(3)Based on neural network model,regional bus travel characteristics and land use planning,a three-stage short-term prediction model for accessibility of transportation system is proposed.The first stage of the model is based on the neural network model,takes the passenger’s bus travel in the historical period as the input,obtains the output of the neural network model,and obtains the bus travel group in the prediction period combined with the designed bus travel rate function.In the second stage,based on the passenger historical travel OD(origin destination)information,the selection probability of the travel group identified in the first stage to the destination is calculated.In the third stage,combined with the land information and the results of the first two stages,the personal accessibility is calculated to obtain the prediction results of the personal accessibility of the public transport system.(4)Taking Chaoyang District,Beijing,China as the research area,the performance of the proposed three-stage prediction model is evaluated.Different data set screening and arrangement schemes are designed,and the accurate value,recall value and F1 value of evaluation indexes are selected to obtain the optimal performance of neural network,so as to ensure the accuracy of short-term prediction of personal accessibility of public transport system.The visual comparison between the predicted value and the real value of accessibility shows that the three-stage short-term prediction model proposed in this paper can accurately predict the spatial-temporal distribution of personal accessibility of public transport system on weekdays and weekends.(5)Based on the space-time constraints of travelers’ activities,a joint activity accessibility evaluation model considering the uncertainty of travel time is proposed.The comprehensive effects of travel time uncertainty,travelers’ risk attitude,fixed location in the travel activity,minimum duration of activities,number of participants in joint activities and other factors on the accessibility of joint activities are considered.