| Due to the rapidity and punctuality of urban rail transit,city managers have taken urban rail transit construction as an important means to ease urban traffic congestion.However,after the completion of urban rail transit,the lines are fixed and cannot be arbitrarily changed.The shortcomings such as poor accessibility have led to inconvenience for passengers traveling in the outskirts of the country,thereby reducing the efficiency of public transport services.The regular bus has the advantages of flexible lines,wide coverage,good accessibility,etc.It is an important means of transport connecting urban rail transit.Therefore,from the point of view of constructing a comprehensive transportation system in which multiple modes of transportation are coordinated with each other,it is necessary to analyze the transit demand of public transport and urban rail transit,so as to optimize the efficiency of rail transit transfer and improve the quality of public transport services.The development of traffic big data and the improvement of data collection technologies have provided data and technical support for the scientific research on residents’ travel behavior characteristics,travel mode choices,and public transportation transfer requirements.This paper uses Shenzhen public transport IC card swiping data,urban rail transit AFC data and public transport GPS data,and draws on domestic and foreign research results to carry out the following research:Firstly,analyze the basic data structure of public transport IC card data,track AFC data,and bus-to-station GPS data;clean raw data by missing value compensation,GPS coordinate offset correction,abnormal data rejection,etc.;Match bus IC card swiping time with bus transit time,identify passenger bus pick-up point location,and use commuter passengers’ fixed travel space-time characteristics to calculate the bus detour point location;Calculate bus and rail traffic passenger OD matrix,identify residents’ travel behavior characteristics,and travel mode selection.Secondly,identify commuting passengers with fixed transfer lines from travelers with characteristics of transfer behavior.Through the cumulative transfer time distribution of this group of passengers,a reasonable transfer time threshold range is used to identify the transfer behavior in the original data;Based on the bus-subway transfer data,the station transfer demand and the characteristics of the bus connection radiation area were analyzed,the parameters of the radiation area were selected,and the selected site classification feature variables were reduced using principal component analysis,and use the K-means algorithm to classify the connected sites.Thirdly,based on the classified sites,the factors affecting the transfer efficiency are analyzed,and from the aspects of high efficiency,comfort and quickness,the evaluation indicators of site transfer efficiency and the evaluation system are selected.Finally,the gray correlation evaluation method is used to evaluate the transfer efficiency of each type of site,and corresponding suggestions for improving the station transfer efficiency are given. |