Analysis Of Spatial Distribution And Demand Characteristics Of Traffic Travel Based On Cruising Cars And Online Car-Hailing Data

Traditional traffic travel surveys are difficult to accurately reveal the characteristics of urban traffic travel due to the limited amount of data,low accuracy,and insufficient sample representativeness.In recent years,with the great development of spatio-temporal big data analysis and processing methods and technologies,it has provided theoretical support for further detailed analysis and characterization of urban traffic travel characteristics.Based on the large-scale spatio-temporal data of cruise cars and online car-hailing,this paper conducts an in-depth study of the spatial distribution of urban traffic travel and the integrity,agglomeration,and structure of travel characteristics,with a view to providing theoretical research and practical management of urban traffic A traffic travel demand analysis method that takes large-scale spatio-temporal data as the research object can be used for reference and reference.The specific research content of this paper includes the following four aspects:(1)The overall study of urban traffic travel demand.The GPS trajectory data and point of interest data of cruise cars and online car-hailing in Xiamen are selected as basic data,and the necessity and feasibility of the integration of the two are explained based on the difference in the temporal and spatial distribution characteristics of cruise car and online car-hailing data;Data preprocessing such as information extraction and aggregation,data cleaning,reclassification,and map matching were performed on trajectory data and point of interest data,and more than 4.8 million travel records and more than 140,000 points of interest records were generated;further,Using nuclear density analysis,Pearson’s coefficient and other methods,from the perspective of time and space,the overall distribution characteristics of travel demand and important points of interest that affect traffic travel are obtained.Laid a data foundation and a holistic research perspective for the characterization of traffic travel and the impact analysis of land use properties below.(2)Spatial distribution analysis of traffic travel based on community discovery algorithm.This paper proposes a framework for dividing traffic travel space based on graph segmentation.The framework establishes a 500-meter grid model,constructs a network topology map with travel volume as the boundary value,and uses the Fast Unfolding community based on modularity.It is found that the algorithm performs graph segmentation;in view of the difference in travel distance,the evolution process of the spatial distribution of traffic travel under the influence of different travel distances is explored,revealing that the spatial distribution of traffic travel has multi-level characteristics,and 18 first-level sub-regions and 5 are obtained.Second-level sub-area:A comparative analysis of the degree of matching between the results of the traffic travel space division and the administrative divisions shows that the traffic travel weakens the influence of the administrative division boundaries.(3)Analysis of the agglomeration of urban transportation travel demand based on complex network theory.According to the spatial analysis unit proposed in the previous study,the local Moran index method of spatial correlation analysis is applied to judge and identify the hotspot areas in the sub-area based on the number of travels starting and ending points;for the existing research,only the traffic hotspot area is identified However,it lacks the insufficiency of detailed and deep research on the characteristics of the agglomeration centers of hotspots.Using the Indicators of Centrality theory of complex networks,the OD amount from hotspots to other non-hotspots and the path between any two non-hotspots are proposed and defined.The core indicators such as the ratio of the OD of the hotspot to the direct OD quantitatively describe the characteristics of the agglomeration centers of the hotspots;further combining the land use properties of the hotspots,the analysis of different land use such as commercial districts,transportation hubs,residential areas and office buildings is analyzed.The interaction and influence of the characteristics of the hotspot cluster center,and further pointed out that the degree of spatial connection between the sub-areas in Xiamen Island is relatively strong,and the hotspot area outside Xiamen Island has a greater impact on the amount of travel in the sub-areas.(4)Structural analysis of urban transportation travel demand based on non-negative matrix factorization algorithm.Considering the high-dimensional complexity and spatiotemporal heterogeneity presented by the travel demand among sub-regions,a nonnegative matrix factorization algorithm is used to reduce the dimensionality of the spatiotemporal OD matrix,revealing the structural characteristics of traffic travel;accordingly,the sub-regions are obtained.There are three main types of travel demand between regions: morning peak travel demand,evening peak travel demand and early morning peak travel demand;a comparison with the data set of previous years shows that the cosine similarity index of the two reaches 95.06%,which verifies the stability of the travel demand structure Further,an optimization model was established to evaluate the nature of its land use.The dominant land use types of the three travel types are commercial service facility land,public management and public facility service land,and residential land.There are 55 figures,11 tables,24 formulas,and 103 references.