A data-driven epidemic model to analyse the course of covid-19 in the Veneto region

The current COVID-19 pandemic is an unprecedented global health crisis, with severe economic impacts and social damages. Mathematical models are playing an important role in this ongoing emergency, providing scientific support to inform public policies worldwide. In this thesis work, an epidemic model for the spread of the novel Coronavirus disease in the Veneto region has been proposed. Starting from the available local Health System data to examine past year contagion numbers and other features potentiality, a SEIQRD (Susceptible Exposed Infected Quarantined Removed Deceased) compartmental schema has been designed generalizing the classic SIR model. Then, the infection dynamics have been practically implemented in two versions: as a Deterministic Equation-based formulation and as an Agent-based model. While the former has been maintained simple and computationally inexpensive in order to serve as a baseline and to quickly provide parameter estimates, for the latter a detailed metapopulation of agents with personalized attributes and network of contacts has been developed to recreate as realistic as possible simulations. Once these models have been trained and validated, they could became valuable tools for various types of analysis and predictions. In particular, the agent-based version, thanks to its flexibility as well as to its higher resolution, could be exploited for exclusive a posteriori evaluations of the effectiveness of the adopted containment measures in reducing the pandemic in Veneto.