Research On The Travel Of Residents,Passenger Flow And Built Environment Impact Based On Multi-Source Data

With the advancement of rail transit construction,the scale of rail transit network is expanding,and the comprehensive optimization needs among residents’ travel demand,rail transit network structure and urban land use are increasingly prominent.At the same time,with the development of traffic big data research in recent years,traffic data with large rules and long time series provide data basis for studying residents’ travel needs.Based on multi-source big data of residents’ travel trajectory,this paper uses trajectory clustering and Arc GIS visualization analysis to analyze residents’ travel characteristics and influencing factors.On this basis,the geographical weighted regression model and the coupling coordination model are constructed to explore the influence of the built environment around the station on the passenger flow of rail transit.The research results play a guiding role in urban comprehensive traffic planning and passenger flow guidance,and create a new pattern of urban development.The main research contents are as follows :(1)Cleaning and deep processing of multi-source data.The GPS data of taxis are matched based on geometric analysis;the location and time of passengers on and off the train are extracted based on the passenger state;for Geolife data,the algorithm based on distinguishing strategy is used to extract stay point data;python was used to crawl POI data representing land use types,and Arc GIS was used to screen the number of various POIs within 800 m of subway station.(2)Based on the deep mining of big data,the spatial and temporal characteristics of residents’ travel are obtained and analyzed.DBSCAN and K-medoids clustering algorithms are used to analyze the travel hotspots of individuals and groups.Inverse Distance Weight and Standard Deviation Elliptic Analysis are used to analyze the change of taxi travel demand from the perspective of space.The temporal distribution of rail transit passenger flow and the flow direction of passenger flow between lines and districts are analyzed.Arc GIS is used to visualize the spatial distribution of passenger flow in different time dimensions.Finally,based on the travel characteristics index,the commuter passenger data accounting for 46.3 % of the passengers studied were screened by Kmeans clustering.Comparative analysis shows that the residents’ travel patterns and spatial-temporal distribution characteristics of the three data are relatively consistent.(3)The GWR model is constructed to analyze the influencing factors of rail transit passenger flow.Twelve factors were selected from the ’5D’ dimensions of density,diversity,design,distance from bus station and destination accessibility as built-up environmental indicators,and geographically weighted regression models were established for rail transit passenger flow in different periods.The visual analysis of regression results was carried out by Arc GIS,and the decisive factors affecting subway passenger flow were determined as the number of households,the number of companies,the density of road network and the distance between adjacent rail transit stations.(4)The coupling relationship between rail transit passenger flow and built environment around the station is studied.Aiming at the separation phenomenon of occupation and residence defined by the ratio of occupation to residence around the station extracted from the OD of commuter passengers,the coupling coordination degree model based on entropy method is used to analyze the coupling coordination degree of rail transit passenger flow and the surrounding built environment.The results show that the proportion of stations with high coupling coordination and above is 57.4 %,and mainly concentrated in the urban core area.Among the stations with low-medium coupling coordination,most of the stations are located in the periphery of the city,and their built environment lags behind the current passenger flow,of which 76.2 % of the station ’ s job-housing ratio is less than 0.8.It is found that increasing the number of households,companies and road network density can effectively improve the coupling degree of branch stations.For stations with low coupling coordination in urban core areas,increasing the number of companies and the number of shopping and entertainment is more effective to improve the coupling coordination.Finally,according to the analysis results of coupling coordination degree,suggestions are put forward for urban land planning and track network optimization.