Identifying Taxi Commuting Traffic Analysis Zones Using Massive GPS Data

Abstract

AbstractThe rapid urbanization process leads to the spatial separation of residence and workplace, which further complicates the commuting pattern of urban residents. Commuting travel by taxi between residence and workplace is convenient and common, while few studies focus on mining commuting pattern using taxi GPS data. In this paper, we propose a taxi commuting traffic analysis zones (TAZs) identification method to identify the regional-level commuting patterns of taxis. The method mainly includes three steps: dividing TAZs considering Point of interest (POI) information, obtaining flow transfer matrix, and identifying commuting TAZs using K-means algorithm. Extensive experiments are conducted on taxi GPS dataset from Chongqing, China. The results show that the method is efficient. 52 pairs of TAZs with commuting relationship are successfully identified, and some typical commuting features are analyzed. The analysis results provide valuable reference for relevant departments and companies, for example, designing custom buses.KeywordsTaxi GPS dataRegional-level commuting patternTAZsPOI