COVID-19 county similarity is a measure of how comparable counties are within a state, with regards to recent COVID-19 risk.
COVID-19 county similarity is calculated daily, by examining the estimated risk exposure (see our methodology) over the last 30 days (see dendrogram matrix below). We then use a K-means algorithm in conjunction with both a silhouette and a gap statistic function, to identify the optimal number of clusters. We then segregate using hierarchical clustering. The groupings identified in the matrix below are visualized in the map.