Authors: Zeyu Zhao, Jiayu Chen, Helen X.H. Bao, Yuhan Zheng, Zhaoyi Li, John E. Taylor, Tianguang Meng, Dongping Fang
Year: 2026
Status: Work in Progress
PDF: Available upon request — hxb20@cam.ac.uk
Research line: AI and computational methods for spatial analysis
Abstract
Conventional resilience metrics prioritise physical infrastructure restoration on the assumption that technical service recovery equates to societal recovery. This study uses a catastrophic historic rainstorm as a natural experiment to quantify the resilience perception gap. By synthesizing millions of geotagged citizen appeals with real-time IoT infrastructure logs, the authors show the gap is structured (not noise) and driven by urbanisation inequalities and the hierarchy of human needs. Physical metrics overestimate resilience for survival necessities (30.2% lag) while underestimating livelihood needs (10.4% surplus).