The built environment is confronted with mounting threats from natural hazards. As built environments grow increasingly complex, such hazards not only inflict damage on physical components including buildings and infrastructure, but also extend their adverse impacts to social dimensions such as residents’ health, safety and overall well-being. This presentation proposes a human‑centric, data‑driven framework for investigating urban systems and addressing their resilience challenges.
Metrics and quantitative methods are developed to evaluate residents’ resilience needs from a human‑centric perspective. Computational models are applied to capture the dynamic evolution of residents’ needs and urban functional performance in the aftermath of rainfall‑related hazards.
A spatial machine learning model for resilience assessment is established. By incorporating spatial heterogeneity with interpret-able analytics, and a Geographically Weighted Random Forest model is proposed to characterise localised spatial relationships.
A hypernetwork simulation framework is developed for post‑disaster recovery analysis. Using a hypernetwork structure, hyperedges are introduced to represent complex interactions among social systems, physical infrastructure and human needs. This supports a holistic understanding of post‑disaster recovery processes within a human‑centric paradigm.
Case studies are presented to validate the feasibility and effectiveness of the proposed methodologies in analyzing and enhancing built environment resilience.