Neighborhood Level Disease Dynamics Using a Metapopulation Model and Cellphone Data

Keywords: disease dynamics, metapopulation, neighborhood patterns, cellphone data
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 03:40 PM (Eastern Time (US & Canada)) - 2/25/2022 05:00 PM (Eastern Time (US & Canada))
Room: Virtual 64

Abstract

People visit multiple places during the day (e.g., for work, school, or ancillary purposes such as shopping, recreation, or services) and make contacts that could facilitate disease transmission among them. Most existing epidemic models fail to capture such spatial and temporal heterogeneities in population mobility. Metapopulation models are commonly the first solution to consider spatial heterogeneity in epidemic models, which suffer from two main shortcomings. First, they have been limited by the availability of population flow data between metapopulations. Second, they ignore temporal heterogeneity in population mobility. This research set out to tackle these issues by using cell phone neighborhood patterns and device home areas from SafeGraph. Neighborhood patterns are collected from individual visits to points of interest in an urban area. Using these data, we generate hourly flow matrices between census tracts in Cook County, IL, for the first week of July 2018. We develop and implement a susceptible, infectious, and recovered (SIR) model that leverages these flow matrices to capture spatiotemporal heterogeneity of population mobility. Our preliminary results show that the proposed model can effectively capture the spatial and temporal heterogeneity of population mobility and their effects on the spread of the disease.