Paris as a 15-Minute City: An Explainable AI Perspective

Service availability matters, but inequality persists: The study found that higher availability of nearby services (shops, amenities, etc.) is associated with less private car use and more walking/cycling in Paris. However, this positive relationship is "substantially weaker in the outer agglomeration," suggesting that the 15-minute city concept's benefits may not reach all residents equally.

Multiple factors beyond urban design influence mobility choices: Machine learning analysis identified six key predictors of travel behavior: trip purpose, home-work distance, local service availability, vehicle ownership, public-transport subscription, and sociodemographic context. This means that simply building services nearby won't change behavior without addressing car ownership and transit access—highlighting the complexity of sustainable mobility policy.

Car ownership is a powerful predictor of car use: The research found that vehicle ownership and driving-license availability are "associated with higher predicted car use" even for short trips where services are nearby. This suggests that reducing car dependence requires addressing not just urban design but also the availability and affordability of alternatives.

Public transit subscription can reduce car dependence in underserved areas: In neighborhoods with sparse services, having a public-transport subscription was "associated with lower predicted car dependence," indicating that transit access may be particularly important for residents in less dense areas who cannot rely on walkability alone.

Explainable AI reveals policy-relevant insights: By using interpretable machine-learning methods rather than black-box models, researchers identified which factors actually drive mobility decisions, enabling more targeted and evidence-based urban planning interventions.