Research On The Influence Of Built Environment On Subway Commuting Behavior In Traffic Analysis Zones Driven By Multi-source Data

Urban rail transit,especially subway systems,are an effective solution for green,smart,and sustainable urban development.The construction and operation of subway systems improve the commuting efficiency and range of activities for urban residents.However,the expansion of subway systems has further exacerbated the degree of separation between work and residential areas and led to an increase in the commuting distance of individual residents.Therefore,it is necessary to explore the interaction between urban built environment factors and residents’ subway travel behavior from a micro-individual perspective,as well as the influencing mechanisms on residents’ choice of subway commuting,in order to optimize the configuration of the built environment for subway travel,and improve the adaptability and share ratio of subway transportation and built environment,and achieve green and low-carbon travel.Taking Xi’an city as the research object,this paper adopts the research approach of "identifying problems,mining data,dividing research units,constructing models,and analyzing independent factors" to analyze the heterogeneous effects of built environment factors on subway passenger flow and reveal the degree of influence of significant built environment factors on residents’ subway travel outcomes.The main research content includes the following four points:(1)Using POI(Point of Interest),subway card swiping,and resident travel survey data as data sources,analyze the spatial distribution characteristics of land use and subway passenger flow in the research area,and obtain the relationship between built environment and subway passenger flow distribution from the data level;(2)Taking urban traffic analysis zones as the starting point,a subway topology network is constructed based on the two-level partition theory.Firstly,the Louvain community detection algorithm is used to divide the research area into transportation central areas.Then,the K-center clustering PAM(Partitioning Around Medoids)algorithm is used to divide the transportation central areas into traffic analysis zones,and the values of various built environment factors in the traffic analysis zones are obtained;(3)Based on the basic difference in data spatial position,a GWR(Geographically Weighted Regression)model is constructed based on the traditional least squares method.Variables such as land use mix,parking density,and intersection density are included to analyze the heterogeneous effects of significant built environment variables on subway passenger flow;(4)Taking micro-individual commuting behavior as the analysis object,a survey of Xi’an residents’ subway commuting behavior is conducted.The Light GBM(Light Gradient Boosting Machine)model of machine learning is used to analyze the effects of personal socioeconomic attributes,built environment attributes,and travel attributes on subway commuting choices,and to reveal the contribution of different characteristic factors to the choice results.The research results of this paper can explore the heterogeneity of urban built environment factors and subway passenger flow in traffic analysis zones and its causes,reveal the contribution of built environment factors to individual travel mode choice results,and provide scientific reference for improving regional built environment,increasing subway commuting population coverage,and alleviating urban traffic congestion.