Urban Jobs-Housing Zone Division Based on Mobile Phone Data

Abstract

Most of the existing researches were divided jobs-housing zones based on temporal activity variation, which were lack of mining spatio-temporal interaction characteristics. With the trend of big data and artificial intelligence, mobile phone data is provided an emerging source for urban research. This paper is proposed traffic semantic concept to extract commuters’ origins and destinations. According to extracted data, four characteristic indexes (including the volumes of user, aggregation, dissipation and new increment) are analyzed traffic semantic attribute. Combining with the geographic information of base stations and traffic semantic, an unsupervised k-means clustering algorithm based on weighted Mahalanobis distance function is used to divide 200 jobs-housing zones in Shenzhen. Moreover, the commuting index is calculated to measure tendency of jobs-housing zones. Compared with the actual land use data, the results are verified reliability of method. All these findings can be helpful to analyze travel behaviors and make urban planning.