Identify Charactertistics Of Rail Stations And Its Influencing Factors

Large-capacity urban rail transit systems play an important role in addressing urban transport issues while also changing and shaping urban form at the micro and macro levels,serving as a window into potential urban change.Among them,the behavioral patterns of passengers at rail transit stations,the characteristics of the urban communities they serve and their evolution,became a research hotspot.However,the traditional data acquisition method has many problems in terms of data completeness,timeliness and so on,which constrains the identification of the charcteristics of the rail transit station and the mining of related factors.In this paper,factors related to station charcteristics are analyzed using data mining methods to identify station charcteristics based on continuous multi-source built environment data in the vicinity of the station,based on rail transit AFC(Auto Fare Collection,AFC)data for multiple consecutive working days.Specifically,the study accomplished the following:Firstly,based on Beijing's rail transit AFC data for five consecutive working days in 2017,the rail transit passenger flow pattern was analyzed.Firstly,the distribution of passenger travel time across the whole network,the average duration of hourly passenger travel,the radius of passenger travel and the average number of trips are studied;secondly,by analyzing the macroscopic time series data of the current situation of the typical station,the morning and evening peak hours are selected as the typical time periods to distinguish the characteristics of station passenger travel;finally,the circular distribution law of the OD of passenger travel throughout the day and during the early peak hours is analyzed.Secondly,passenger travel characteristics were quantified by constructing the spatio-temporal probability of passenger travel.The Gaussian process mixed model was used to analyze the clustering of stations based on the above-mentioned indicators,and five typical station service functions were extracted,including multimodal interchange hub and leisure cluster;residential cluster;employment cluster;mixed but mainly residential cluster;and a mixed residential and employment cluster.Thirdly,based on the data of built environment around the station,the level of various built environment within 500 meters of the station is statistically analyzed,including the spatial location of the station,information on the surrounding housing,information on the surrounding food and beverage,information on tourist attractions,public transportation,educational institutions and other POI information;the correlation between the built environment factors and the station service function within 500 meters of the station is analyzed,and factors with significant relationship with the station service function are identified based on the Multinominal Logit Model.The paper firstly quantifies the travel characteristics of passengers by calculating the spatial-temporal travel probability of passengers based on the Bayesian framework,which makes it possible to consider high-frequency passengers and low-frequency passengers on the same scale.Secondly,it integrates the passenger flow characteristics at the macro and micro levels,digs up the typical charcteristics of rail transit stations,and improves the interpretability of the identification results of station charcteristics by using the personal travel behavior of passengers and the macro passenger flow laws of stations for mutual verification.Finally,by analyzing the correlation relationship between station charcteristics and built environment factors,it shows that there is a significant correlation relationship among station spatial-temporal characteristics,station built environment and station charcteristics.