Between Real and Imagined: Developing a dynamic immersive virtual auditing (DIVA) framework for high-density urban health trails

High-density urban environments present significant challenges for designing public spaces that promote health and well-being. Although previous studies have employed virtual reality to investigate human–environment interactions in urban contexts, most have focused on static, low-fidelity, subjective and qualitative approaches. To address this gap, this paper introduces Dynamic Immersive Virtual Auditing (DIVA) framework that integrates data-driven analyses into the evaluation of health trails in a realistic, dynamic way. The framework utilizes high-fidelity 3D reconstruction to enhance simulation realism and systematically modifies baseline models to test spatial elements. Participants interact with those immersive, continuous virtual environments with free-viewing capabilities, while real-time physiological and self-reported data are collected. ANOVA is employed to identify significant effects of spatial elements on psychophysiological response, while machine learning uncovers nonlinear relationships among design factors, offering deeper insights into optimal restorative configurations. Preliminary frameworks suggest that this phygital paradigm can support evidence-based design methodologies, validate design decisions, and provide theoretical contributions of urban health design.