Detecting And Analyzing Urban Centers Using Taxi Trajectory Data

As the key to understanding urban structure,the identification and analysis of urban centers’ geographical locations and attributes have been widely concerned by researchers.Due to the close correlation between urban centers and human economic and social activities,many kinds of geographic big data(e.g.,nighttime light data,etc.),which can reflect the trajectory of human activities,are used to extract and analyze urban centers.The nighttime light data can only reflect human activities at night,and its temporal and spatial resolutions are limited by remote sensing satellites.Compared with nighttime light data,taxi trajectory data has flexible temporal and spatial resolutions and can reflect human activities throughout a 24-hour period.In addition,this paper believes that the relationship between taxi trajectory data and human activities is much closer.Therefore,this paper aimed to explore the feasibility of taxi trajectory data in the identification and analysis of urban centers and the advantages compared with nighttime light data.The main work of this paper is as follows:(1)Based on the existing research work,this paper summarized the limitations of using nighttime light data to analyze urban centers,put forward the idea of using taxi trajectory data to replace nighttime light data in urban center extraction and designed the analysis framework of detecting and analyzing urban centers using taxi trajectory data.(2)This paper took Shanghai as the research area to validate the analysis framework,extracted pick-up and drop-off spots from taxi trajectory data,and drew a statistical grid with a side length of 500 meters to calculate the hourly average number of pick-up and drop-off spots in each grid.In this paper,the statistical grid was considered as topography and contour lines were generated from it.The localized contour tree method was applied to construct and simplify contour trees,and finally,the boundary and hierarchical structure of urban centers were identified.(3)This paper classified the detected urban centers as two types: Monocentric Urban Center(MUC)and Polycentric Urban Center(PUC).Five statistical attributes,four morphological attributes,and intensity gradient was designed for MUC,and the calculation of polycentricity was designed for PUC.In addition,this paper made comparisons between weekdays and weekends,and between Monday and Sunday(7:00-10:00)in order to explore the spatial and temporal characteristics of urban centers.The experimental results of this paper showed that the taxi trajectory data has the advantages of high sampling density and flexible spatial and temporal resolution in the identification and analysis of urban centers.Compared with the results of nighttime light data,the experimental results of this paper were more detailed,and more urban centers were obtained.Moreover,the spatial and temporal characteristics of urban centers can be analyzed through the statistical grids of different time periods.This paper combined the remote sensing method with taxi trajectory data,which provided a new perspective for the analysis of vector data including taxi trajectory data,and complemented the existing urban center analysis.