Abstract In Japan, railway operators have played a pivotal role in promoting urban mobility and consumption by managing commercial facilities in conjunction with transportation infrastructure based on population growth. However, the population has declined in recent years. It is imperative to leverage existing infrastructure to its maximum potential to ensure the continued retention of current users and attract new users. The objective of this study was to elucidate the visitation and consumption patterns surrounding railway stations by integrating the analysis of transportation and purchasing behavior, which have traditionally been examined independently. In this study, two independent datasets were utilized: transit smart card data for ticket gate entry and exit and loyalty card data from commercial facilities in the station integrated area. The utilization of these datasets makes it possible to estimate the number of individuals residing in the facilities and surrounding areas. A seven-dimensional tensor was constructed, incorporating five variables: “Time Slot,” “Weekday/Holiday,” “Gender,” “Age Group,” “Home-to-Station Distance,”, and two indices: “PSP,” “P/T Category.” The remaining populations from both data sources were aggregated, and nonnegative tensor factorization was applied to extract latent behavioral patterns. Consequently, policy-relevant user segments were identified, including females who made purchases during the weekday daytime. The significance of this study lies in the development of an integrated method for analyzing transportation and purchasing behaviors using mutually independent datasets. The methodology of this study establishes a framework for conducting further in-depth analysis using currently available big data.
Sasaki et al. (Wed,) studied this question.