HPWO model: A housing-price weighted opportunities model for heterogeneous human mobility

Predicting urban human mobility is a vital topic in urban planning, transportation engineering, and public health management. Understanding universal patterns within complex travel distributions is crucial for improving the interpretability and practical applicability of mobility models. However, most existing models rely on demographic data as primary metrics and overlook the heterogeneity in travel generation and attraction, making it difficult to model the diversity of population movements. To address this challenge, this study employs machine learning and spatial statistical methods to uncover the relationship between housing prices and travel behavior. Inspired by intervening opportunities theory, it proposes the Housing-Price Weighted Opportunities (HPWO) model, which utilizes housing prices to characterize travel dynamics: origin housing prices capture travelers' preferences, destination housing prices reflect attractiveness, and network-based intervening housing price distance represents impedance. A case study using metro commuting data from three Chinese cities demonstrates that the proposed model achieves strong predictive performance in terms of travel volume, arrival volume, and spatial distribution. Furthermore, it effectively captures the diversity of human mobility within urban systems. This study offers novel insights for modeling heterogeneous travel distribution.