Abstract: This study investigates the impact of human mobility on the spatio-temporal dynamics of COVID-19 spread by utilizing daily COVID-19 data of more than 300 cities in China. The spread of COVID-19 in China is characterized by a two-stage pattern, namely the inter-city transmission driven by human mobility in the first stage and the local transmission among family members in the second stage, which have further shaped the spatio-temporal patterns. The impacts of human mobility on the COVID-19 spread are featured by structural differences and dynamic patterns. Temporal movements including tourism and business travel are the main route of transmission at the beginning of COVID-19 outbreak, while internal migrants returning to their hometowns during the Chinese Spring Festival are mainly responsible for the peak outbreak in early February. Both survey data and big data have equally high statistical power in predicting and interpreting the spread of COVID-19, indicating that the combination of the twos strengths would contribute substantially to advancing quantitative research and improving social governance capacity.

Keywords: COVID-19, Human Mobility, Spatio-temporal Dynamics, Big Data

摘要： 基于城市层面的每日疫情数据，利用空间可视化和统计分析模型，从人口流动视角考察中国新冠疫情扩散的时空动态规律，探究不同类型人口流动的影响。研究发现，人口流动带来的城际传播和家庭为主的本地传播构成了中国疫情扩散的两阶段模式，塑造了疫情的时空格局。人口流动对疫情传播的影响具有结构性差异和动态性特征，商务、旅游等短期人口流动是疫情暴发初期的主要传播途径，长期人口流动带来的春节返乡流则推动了疫情进入高峰期。医疗资源并未对疫情防控产生系统性约束，人口老龄化的提高也未导致城市感染人数增加。传统数据和大数据对疫情传播的解释和预测具有同等效力，二者优势的结合是深化定量社会研究、提升社会治理能力的有效途径。

关键词: 新冠疫情, 人口流动, 时空动态, 大数据