Data-Driven Travel Analysis for Mobility Tracking

The rise of continuously and passively collected movement data presents fresh opportunities to examine the long-term evolution of personal travel habits from a data-driven standpoint.This research introduces a framework based on clustering to pinpoint travel behaviour patterns and identify possible periods of change at the individual level.Initially, we isolate key trips that illustrate distinctive individual movements.Next, by considering trip mode, distance, and duration dimensions of travel behaviour, we assess the similarities among trips and categorize them into clusters through hierarchical clustering techniques.These trip clusters symbolize different aspects of travel behaviours, and the variations in their relative proportions over time indicate shifts in travel preferences.We employ two methods to identify changes in travel behaviour patterns: one based on the Herfindahl-Hirschman index and another utilising a sliding window approach.The framework is validated with data from a comprehensive longitudinal GPS tracking study where participants were offered Mobilityas-a-Service (MaaS).The methods effectively reveal notable changes in travel behaviour among users.Additionally, we examine how the MaaS offering influences individual travel behaviours based on the detected changes.The suggested framework for detecting behavioural changes offers significant insights for managing