Exploring space–behavior interactions during walking in nonlinear architectural spaces based on multi-source data

With the rapid development of parametric design and construction technologies, contemporary nonlinear architectural spaces are reshaping people’s walking behaviors and experiences within them. This study focuses on the Changsha Meixihu Art Museum, a representative nonlinear architectural space. Based on on-site walking experiments with 28 participants, it integrates multi-source data from visibility graph analysis (VGA), video-stimulated recall semi-structured interviews (VSRI), and heart‑rate variability (HRV) to explore space–behavior interactions during walking in nonlinear architectural spaces. The integrated analysis of multisource data indicates that incomplete spatial information resulting from path bifurcations and visual occlusion in nonlinear architectural spaces does not necessarily diminish experiential quality. Instead, this incomplete information and the visual attraction arising from curved interfaces, spatial scale, and variations in light and shadow jointly constitute important conditions shaping participants’ decisions to pause and observe, actively explore, and adjust their walking behavior. This finding indicates that the optimization of nonlinear architectural spaces and their associated behavioral interaction experiences should not aim to eliminate spatial uncertainty entirely, but rather to achieve an appropriate balance between spatial legibility and exploratory potential. This study, through the integration of multi-source data, provides insights for enhancing the design and optimization of experience-oriented nonlinear architectural spaces.