| With the rapid development of China’s economy and society and the deepening of the urbanization process,the number of urban motor vehicles is also growing rapidly,but it is limited by the limited space resources of urban road traffic and the limitations of investment in road planning and construction.The development of transportation lags behind the rapid growth of motor vehicles,making the pressure of urban transportation increasingly serious,and the problem of urban transportation roads has become more and more obvious.Based on the analysis of urban taxi GPS trajectory data,the spatial and temporal analysis can identify the characteristics of urban traffic,and by detecting its correlation with the road network structure,it can effectively relieve the pressure of urban traffic and provide basic support for optimizing urban traffic planning and management.This article uses linear streets as the analysis unit to perform data preprocessing operations such as removing noise points,redundancy and map matching on the GPS trajectory data of 14,000 taxis within the third ring road of Chengdu from August 18 to August 24,2014.The statistics on the number of passengers on and off the streets of Chengdu and the dynamic changes are used to explore the spatial and temporal characteristics of urban taxi traffic,cluster the streets according to the Bisect K-means method,and then use the urban road network traffic proximity,intermediary,and direct network three central indicators and streets Correlation analysis of internal and external passenger activity density.The main research contents are summarized as follows:(1)Analysis of the spatiotemporal characteristics of taxi trajectory.Using the aggregation method of working days and rest days and analyzing with one hour as the time interval,it is found that the distribution curve of taxi activity time on rest days is smoother than that on working days,and the peak of working day evening peaks from 18:00.The day appears at 19:00.Hotspots for taxi activities mainly include: Chengdu Railway Station,Chengdu East Station,Chunxi Road commercial district,Sichuan University,Wuhou Temple Museum,Sichuan Provincial People’s Hospital and other major blocks.(2)Street type detection based on spatiotemporal features of taxi trajectory.Using the improved Bisect K-means clustering method,the initial street types are classified according to the street function VG vector to generate four first-level road types of A(close to downtown),B(close cultural district),C(nearly hub area),and D(near residential areas).Among them,A-type streets are mainly distributed within the second ring road.There is a clear peak during peak hours,and there are far more drop-off activities than drop-off activities.During the evening peak period,there is a trend of more drop-off activities than drop-off activities.The three peak periods of the B-type street are more obvious.The morning peak rushing out activity is much more than the drop-off activity.During the afternoon and evening rush hours,the drop-off activity is more than the dropoff activity..The morning and afternoon peaks of the C-type street are relatively gentle,but during the evening peak hours,the passenger activity is more significant,mainly distributed between the Third Ring Road and the Middle Ring Road.D-type streets have an early peak between 7 and 8 o’clock,the afternoon peak is relatively gentle,and the evening peak continues to increase after 17:00.Most of this type of street is distributed near urban residential areas.Then,for each type of primary road type,the clustering is continued according to the street dynamic capacity VR,and three secondary road types are generated.Then,the primary road and the secondary road type are compounded.Due to the active level of A2,C1,and C2 type streets Low,classify it as type E(inactive type),and finally get 10 types of traffic network streets.(3)Correlation analysis of centrality of urban traffic street network and taxi trajectory in Chengdu.Proximity centrality presents a "center-periphery" distribution pattern,intermediary centrality is also mainly distributed within the central urban area,and direct centrality presents a multi-center distribution pattern.By analyzing the characteristics of the density of taxi pick-up and drop-off activities in the street unit and the centrality of the street network,it is found that in the global perspective,the proximity centrality and intermediary centrality and their correlation are higher than the direct centrality;while in the local perspective,three types of centrality indicators The correlations with the density of taxi rides are increasing as the search radius increases.The correlation trends of the network centrality and street activity density of the four street types A,B,C,and D are consistent with the overall street network centrality and activity density.The correlation characteristics of the C-type street are more obvious.Under the radius,the centrality of the Ctype street network is better than other street types.Using the street as the basic analysis unit to analyze the spatiotemporal characteristics of taxi trajectories,revealing the spatiotemporal heterogeneity of urban travel characteristics on different streets,expanding the perspective of GIS-based trajectory data analysis;its correlation analysis with the urban road network structure reflects The influence of the centrality of the street network on urban travel helps to optimize urban transportation planning and refine urban traffic management and control,as well as improve the level of refined urban governance. |
