Measuring Spatial Distribution of Tourist Flows Based on Cellular Signalling Data: A Case Study of Shangha

Tourist flows have significant effects on the urban external transport. This study applied cellular signaling data to identify tourists and introduced the notion of space of flows to depict their spatio-temporal distribution. Firstly, using unlabeled cellular signaling data and POI data, a methodology of tourist identification was proposed. Residents and temporal visitors were distinguished based on data features and artificial rules, and then tourists were screened from visitors based on DBSCAN algorithm. Then, introducing the notion of space of flows, the spatio-temporal distribution of tourists was analyzed. Space of tourist flows was characterized from three aspects: the strength of interaction, the symmetry of interaction and the structure of the network, which was summarized as three S-dimension indices. Finally, a case study of Shanghai was carried out to demonstrate the proposed methodology. As a result, 6.64% of mobile subscribers were identified as tourists. As three S-dimension indices of space of tourist flows were calculated in workdays, holidays and weekends respectively, it was founded that the spatio-temporal distribution of tourists was significantly affected by holidays, high-speed rail routes and the position in the city. The methodology proposed in this paper benefits the understanding of tourist activities beyond traditional survey data and provides valuable references for urban external transport planning at the city scale.