Using GeoAI to examine infectious diseases spread in a hyperdense city: A case study of the 2022 Hong Kong COVID-19 Omicron wave

Abstract

<p>This study utilizes self-organizing maps (SOMs) to investigate the spatiotemporal diffusion patterns and clusters of the 2022 COVID-19 Omicron variant in Hong Kong, incorporating various sociodemographic and environmental datasets. A large dataset necessarily creates a higher dimension in structure, making it challenging for humans to explore the complex associations among many variables and observations. SOMs effectively reduce data dimensions while preserving the topological structure of the original data through unsupervised artificial neural network approaches. We found that many non-centric residential areas repeatedly exhibited similar diffusion patterns over time after the relaxation of anti-pandemic measures. Notably, several non-centric localities with fewer commercial establishments and transportation hubs often became infection and transmission clusters due to temporary increase in crowds during the anti-epidemic measures. Areas with more older housing and industrial facilities were also identified as vulnerable to COVID-19 diffusion due to outdated building structures and equipment. Findings emphasize the need for tailored interventions in local neighborhoods, as well as densely populated commercial and business districts, to effectively manage and prevent infectious diseases in dense urban areas. This study showcases the utility of geospatial AI techniques in analyzing spatial and temporal diffusion patterns of infectious diseases and designing proper measures for their control and prevention.</p>