Research On Traffic Analysis Zones Divition Method Based On Multi-source Data

Traffic Analysis Zones(TAZs)are the basic units of traffic demand forecasting.The proper zoning of Traffic Zones will directly affect the research of future traffic planning.In the era of big data,the traditional traffic district division method is difficult to take into account the current residents' travel characteristics due to its static nature.The result of traffic zoning obtained by combining the actual urban travel big data and land use situation is more beneficial to traffic planning and urban spatial structure analysis.Based on the above background,this paper proposes a multi-source urban big data-driven traffic plot generation method which takes the actual block plot as the basic aggregation unit.First,the clustering features of multiple blocks are mined from the big data of multi-source cities.Mining the land use characteristics of blocks from the POI point of interest data;From the data of public transportation smart card,bike-sharing rental,and taxi and online ride-hailing,the spatial-temporal travel characteristics and transfer travel characteristics of blocks are mined.Based on these data characteristics,a traffic community aggregation algorithm is proposed to aggregate blocks into traffic communities.Firstly,based on the spatio-temporal travel flow characteristics of bus and subway stations,the traffic district aggregation algorithm USES the fuzzy k-means algorithm to cluster the stations,and then selects the core blocks of each potential traffic district based on the clustering results of the stations and the core contribution of the blocks.Then,based on the selected core block as the center,based on the actual land use characteristics of the block driven by POI big data and the time-space travel characteristics excavated from multi-source data such as taxi,online car-hailing and bike-sharing,the block plot is aggregated into a traffic community.A homogeneous loss calculation method is proposed to determine the optimal number of partitions.In this paper,the land nature and spatial-temporal traffic travel characteristics in TAZ are taken as the division target,which effectively solves the problem that the previous division based on big data separated the flow,travel mode and actual land use.In addition,this paper takes the actual block plot as the basic aggregation unit to solve the problem that the previous meshing results do not corres pond to the actual district boundaries and are of poor applicability.Finally,the inner area of Beijing sixth ring road is used as the case study area for the division of traffic districts.By applying the generation method of traffic districts proposed i n this paper and combining with the actual big data of multi-source cities,the division results of 602 traffic districts are obtained,and the results are compared and analyzed with the actual division of Beijing traffic districts.The results show that this method can improve the rationality of traffic planning and urban traffic planning.