Activity Modelling Using Journey Pairing of Taxi Trajectory Data

Taxi GPS data offers an opportunity to discover behavioural patterns in urban populations. However, raw taxi journey data does not provide a link between outbound and return journeys of individual travellers. Without this information, it is not possible to track individual behaviours. In this study, we propose a novel method for pairing taxi journeys and apply it to taxi trajectory data for the city of Shenzhen, China. Journeys related to three activities are considered: shopping, medical, and work. Results, validated using questionnaire data collected in Shenzhen, quantitatively reveal behavioural patterns and suggest possibilities for applications in urban design.