Unfolding and modeling the recovery process after COVID lockdowns

Xuan Yang,Yang Yang,Chenhao Tan,Yinghe Lin,Zhengzhe Fu,Fei Wu,Yueting Zhuang

DOI: https://doi.org/10.1038/s41598-023-30100-5

IF: 4.6

2023-03-13

Scientific Reports

Abstract:Abstract Lockdown is a common policy used to deter the spread of COVID-19. However, the question of how our society comes back to life after a lockdown remains an open one. Understanding how cities bounce back from lockdown is critical for promoting the global economy and preparing for future pandemics. Here, we propose a novel computational method based on electricity data to study the recovery process, and conduct a case study on the city of Hangzhou. With the designed Recovery Index , we find a variety of recovery patterns in main sectors. One of the main reasons for this difference is policy; therefore, we aim to answer the question of how policies can best facilitate the recovery of society. We first analyze how policy affects sectors and employ a change-point detection algorithm to provide a non-subjective approach to policy assessment. Furthermore, we design a model that can predict future recovery, allowing policies to be adjusted accordingly in advance. Specifically, we develop a deep neural network, TPG , to model recovery trends, which utilizes the graph structure learning to perceive influences between sectors. Simulation experiments using our model offer insights for policy-making: the government should prioritize supporting sectors that have greater influence on others and are influential on the whole economy.

multidisciplinary sciences