Study On The Analysis Of Residents' Travel Based On Taxi Trajectory Data

As an important part of urban transportation,taxi has the characteristics of uninterrupted and wide coverage.There is no specific route and station during its operation,which can well reflect the residents' travel behavior,urban traffic operation status.Therefore,more and more scholars use taxi trajectory data to analyze the residents' travel,which provides a scientific and technical means for better revealing the relationship between people,vehicles and roads.Based on the big data of taxi trajectory,combined with the urban road network and POI(point of interest)data,the paper explores the spatial and temporal rules of residents 'travel,and proposes an improved Euclidean distance DBSCAN(Density-Based Spatial Clustering of Applications with Noise)algorithm to deeply explore the hot spot of urban residents' travel.Then,the paper constructs a visualized system of residents' travel analysis,and analyzes the potential relationship between urban spatial structure and residents' travel laws.The specific research content is as follows:(1)Preprocessing of taxi trajectory data and analysis of passenger data.First,the paper studies the structure of taxi trajectory data,urban road network data,and POI data.Aiming at the problems of data errors and redundancy in taxi trajectory data,the paper proposes a method for preprocessing taxi trajectory data.Secondly,the paper performs map matching and correction on the trajectory data that deviates from the road.Finally,due to the temporal and spatial continuity of taxi trajectory data,the paper identifies the taxi trajectory,proposes an analysis model of taxi trajectory data for loading and unloading passengers,and provides a data basis for the analysis of residents' travel time and space and hot spot mining.(2)Study on the analysis of the spatiotemporal characteristics of residents' travel.The paper takes taxi trajectory data as the research object,correlates the urban road network and the urban POI distribution,analyses the distribution characteristics of the trip volume of different date attributes such as working days,non-working days,and Spring Festival holidays in time dimension.The paper studies the distribution of the average travel speed of residents and the time of carrying passengers in different periods of 24 h.the paper analyses the travel locations,travel roads and urban POI distributions in the spatial dimension,and explores the correlation information between the spatiotemporal characteristics of residents' travel behaviors and the spatial structure.(3)Mining of residents' travel hot spot based on clustering algorithm.First,the paper studies the traditional DBSCAN algorithm and data similarity distance measurement method;Second,the traditional DBSCAN algorithm has low calculation efficiency and high memory requirements.The paper considers the road feature information of the trajectory data and proposes a road coefficient to represent the degree of influence of the data road attributes on the similarity between the data in each dimension,and designs a DBSCAN algorithm that uses road coefficients to improve the Euclidean distance.After testing: the improved DBSCAN algorithm is significantly better than the traditional DBSCAN algorithm in terms of clustering accuracy and calculation time.Finally,based on the early peak and late peak taxi load and load data in Shijiazhuang City,the paper uses the improved DBSCAN algorithm to mine hot spot areas of early peak and late peak in Shijiazhuang City,and analyzes the potential relationship between residents' travel behavior and urban resource allocation.(4)This paper develops a visualization system for residents' trip analysis based on taxi trajectory data in Shijiazhuang.The system is designed with a resident trip data management module and a resident trip analysis visualization module,and implements taxi data storage,processing,analysis,and visualization functions.In conclusion,based on the taxi trajectory data of Shijiazhuang,this paper conducts preprocessing and data mining for Shijiazhuang taxi track data,analyzes the travel rules of Shijiazhuang residents.Based on the improved Euclidean distance DBSCAN algorithm of road coefficient,the accuracy and efficiency of hot spot area analysis are improved.Developed a resident travel analysis and visualization system that can provide services for traffic management and traffic planning.