Post-pandemic public transport resilience and mode shift dynamics in India

The COVID-19 pandemic disrupted urban travel, with conflicting perspectives on the permanence of these changes. This study analyses data from 48,839 respondents across twelve diverse Indian cities, using a mixed-methods approach, including machine learning (ML) and Double Machine Learning (DML) to examine pre- and post-pandemic mode choice dynamics at aggregate and city-specific levels. The ML analysis identified fundamental life circumstances as the primary predictors of mode choice. The DML analysis revealed that while public transport (PT) demonstrated significant resilience, powerful behavioural inertia persists, and specific service failures causally deter PT adoption. Pre-pandemic private vehicle use is causally linked to a lower likelihood of shifting to PT. Furthermore, safety and comfort issues, such as station cleanliness and staff professionalism, are causally linked to negative passenger perceptions. The study highlights significant city-specific variations and informs targeted, actionable, evidence-based policy recommendations for developing more resilient and environmentally sustainable urban transport systems.