A Method to Mine Movement Patterns between Zones: A Case Study of Subway Commuters in Shanghai

Abstract

As identifying people's movements across zones can improve our understanding of transportation patterns and recommend strategies for urban planning such as precise locations for targeted advertisements, residential zoning, and transportation development. However, when the amount of data is large or the relationships between data are complex, traditional algorithms for movement patterns between zones become ineffective. We propose a new agglomeration algorithm, namely, the density-based movement patterns between zones (DBMPZ), to mine spatial clustering of movement patterns. To validate the proposed algorithm, we use a real-world dataset of subway commuters in Shanghai and some synthetic datasets to identify movement patterns between zones. The experiment results show that the proposed algorithm can effectively mine movement patterns between zones with high precision, effectiveness, and efficiency. In addition, the proposed algorithm can also play an important role in other regions or types of transportation dataset by modifying the clustering procedure.