Predicting active travel durations in Tehran: A multilayer perceptron approach

Abstract

This study investigates the impact of built environment factors (BEFs) on active travel (walking and cycling) habits in Tehran, Iran, aiming to reduce car dependency and promote sustainable urban transportation. Utilizing data from the Transport, Health, and Environment Survey (THES), the study employs Multilayer Perceptron (MLP) models to analyze how walkability, cyclability, destination accessibility, and land use density influence the duration and frequency of walking and cycling trips across different purposes: shopping, recreational, cultural, and work-related travel. Key findings reveal that walkability and destination accessibility significantly influence walking trip durations, while cyclability and land use density strongly impact cycling trip durations, particularly for retail and leisure activities. Sensitivity analysis demonstrates that even minor improvements (5–20 %) in perceived BEFs can lead to substantial increases in active travel durations. These findings underscore the critical role of urban planning in promoting sustainable transportation, emphasizing the need for policies that enhance walkability, cycling infrastructure, and accessible land use planning to encourage active travel and reduce car dependency in urban areas.