Understanding user demographics using public transportation data

Abstract

As smart cards become dominant in public transportation, more data is collected that capture passenger behaviours: their trip times, which stations do they hop in/out, etc. Understanding passenger demographics can have important applications for health and marketing, as well as public transportation. For example, if we know that a certain metro station has high concentration of young children (students), then the transport authority may increase the security in the station to ensure the children's safety. In this work, we collect real smart card data from a cosmopolitan city in the Middle East and analyse how it relates to the underlying passenger demographics. Our analysis illustrates how association rule-mining can expose rules that are indicative of age, gender and nationality based on usage patterns of public transportation.