Diverse habits, disrupted context and random behaviour in transit demand

Abstract

Public transport systems are central to urban mobility, providing access to jobs, education, and social activities for millions of residents. While conventional models assume rational decision-making, growing evidence shows that travel behaviour is shaped by habits, disruptions, and randomness, creating variability in demand. This thesis examines behavioural variation in public transport use, focusing on heterogeneity, disruption, and randomness.

Habits are analysed as the foundation of travel behaviour, defined by frequency, consistency, and persistence. Using these dimensions, the study develops metrics for system-level habituality, including the number and intensity of habits, their prevalence in journeys, and the share of habitual travellers. Disruptions—systemic, such as the introduction of light rail, and personal, such as residential relocation—reveal how routines are modified. Remaining unexplained variability is treated as randomness, capturing the unpredictable component of demand.

The research uses 8.5 years of smartcard data from Canberra’s MyWay system (2016–2024), enabling detailed analysis of how travel routines form, persist, and change. A persistency index is introduced to infer home locations and detect relocation. Results show that habitual travellers form a stable core of transit demand, underpinning predictability. Relocation highlights sensitivity to personal disruption, while the light rail reshaped demand across specific cohorts and geographies. Irregular users also contribute substantially to overall demand, underscoring the need for systems that remain attractive to infrequent riders.

By foregrounding heterogeneity, disruption, and randomness, this thesis offers a more nuanced understanding of transit demand. It emphasises maintaining reliable services for habitual users while recognising the potential of disruptions to shift behaviour. The findings support the design of more resilient, inclusive, and user-centred transport systems.