Human mobility models for the spread of infectious disease

The transmission patterns of infectious diseases depend crucially on how humans move geographically. While infectious disease data are increasingly becoming available at a fine geographic resolution, our understanding of human mobility at similar scales is relatively lacking. Most mathematical models of geographic disease transmission rely on the so-called `gravity model', developed in the 1940s, that describes the movement of individuals between cities as a heuristic function of the population sizes of their home and destination locations. Despite its widespread use, the gravity model has a number of analytical inconsistencies, and does not adequately account for the true variation in human mobility tendencies. This project will involve both theoretical and data-driven investigation to determine how to better model the movement of humans, with application to the spread of infectious diseases. The first step will be an exploratory analysis of how disease transmission models based on different mobility proxies, including the traditional gravity model, the newer `radiation model', and empirical commuting times and distances based on Google Maps data or similar, might yield different disease transmission patterns. Then, using a state-of-the-art dataset on human movement patterns from the UK collected through a geo-tracking mobile phone application, general mathematical descriptions (`kernels') of the movement tendencies of UK residents will be developed. These kernels will form the basis of a theoretical investigation into how individual movement habits translate into city-level geographic patterns in outbreak timing.