Simulating Epidemics and Interventions on High Resolution Social Networks

Mathematical models of disease spreading are a key factor of ensuring that we are prepared to deal with the next epidemic. They allow us to predict how an infection will spread throughout a population, thereby allowing us to make intelligent choices when attempting to contain the disease. Whether due to a lack of empirical data, a lack of computational power, a lack of biological understanding, or some combination thereof, traditional models must make sweeping assumptions about the behavior of a population during an epidemic. In this thesis, we implement granular epidemic simulations using a rich social network constructed from real-world interactions. We develop computational models for three diseases, and we use these simulations to demonstrate the effects of twelve potential intervention strategies, both before and during a hypothetical epidemic. We show how representing a population as a temporal graph and applying existing graph metrics can lead to more effective interventions