Urban big data analytics for COVID-19 risk prediction and control

Abstract

Prof. Shi and his team have developed a series of extended Weighted Kernel Density Estimation (E-WKDE) models for short-term prediction of COVID-19 symptom onset risk. Compared with most existing COVID-19 risk predictions focusing on confirmed cases, the models can advance the prediction of COVID-19 transmission trend for 4 to 5 days. They were used for evaluating the contribution of the Wuhan lockdown to reducing the onset risk in the rest of China; developing a risk-based vaccine distribution scheme for Hong Kong; studying the transmissibility of different SARS-CoV-2 variants; and evaluating the efficacy of anti-epidemic measures worldwide. Based on the E-WKDE model predictions, the team had submitted over 40 reports to the Hong Kong Government as references for evaluating COVID-19 control measures since early 2020.