Exploring time allocation patterns for daily activities through a discrete choice modeling approach

This study examines human daily mobility patterns by analyzing activity duration choices through discrete choice models, offering insights into travel behaviors across diverse geographic and cultural contexts in the U.S. Using data from the 2023 American Time Use Survey (ATUS), we assessed activity time allocation across six states: California, New York, Pennsylvania, Texas, Illinois, and Florida. By estimating multiple segmented Multinomial Logit Models, we explored how activity types and socio-economic factors influence time allocation decisions. Our findings reveal substantial regional variability in activity durations, with extended periods (120 + minutes) preferred for work and leisure, while brief durations were largely avoided. Dining patterns exhibited more variability, with coastal states favoring moderate durations (30–60 min) and inland states showing more evenly distributed preferences. The pandemic years (2020–2022) prompted notable shifts in work and leisure patterns, particularly in Florida, while dining preferences shifted towards shorter durations, especially in coastal states. This study highlights the importance of context-specific analyses, emphasizing the need for models that capture regional differences in behavior. Our findings suggest significant implications for transportation planning and policy, providing valuable insights for improving travel demand forecasting and facilitating associated infrastructure.