Title: The identification of spatially constrained homogeneous clusters of Covid-19 transmission

Abstract: The paper introduces an approach to identify a set of spatially constrained homogeneous areas maximally homogeneous in terms of epidemic trends. The proposed hierarchical algorithm is based on the Dynamic TimeWarping distances between epidemic time trends where units are constrained by a spatial proximity graph. The paper includes two different applications of this approach to Italy, based on different data (number of positive test and number of differential deaths, with respect to the previous years) and on different observational units (provinces and Labour Market Areas). Both applications, above all the one related to Labour Market Areas, show the existence of well-defined areas, where the dynamics of growth of the infection have been strongly differentiated. The adoption of the same lock-down policy throughout the entire national territory has been therefore sub-optimal, showing once again the urgent need for local data-driven policies.