Dynamic assessment of visitor arrivals: exploring the multidimensional cultural heritage tourism resilience

Tourism resilience has become a central concern amid increasingly frequent compound crises, yet phase-contingent assessments of how socio-spatial factors shape visitor recovery remain limited, particularly for cultural heritage destinations. This study advances a Social–Spatial–Behavioral framework grounded in Lefebvre’s theory of space. Using multi-source spatiotemporal data (2020–2024) from 38 sites in Suzhou Ancient City, China, a sequential approach combining Random Forest, Ordinary Least Squares Regression, and Multi-Scale Geographically Weighted Regression is applied. Results identify four recovery phases and shifting driver salience: social media sentiment dominates uncertainty, transit accessibility shapes constrained recovery, and service diversity supports normalization. Spatial heterogeneity is pronounced across sites. The study offers a dynamic analytical framework and phase-sensitive strategies for heritage destination management.