K-Means Clustering of COVID-19 Cases in Indonesia’s Provinces

Abstract

The novel coronavirus disease (COVID-19) has been rapidly spreading, causing a severe health crisis all around the world, including Indonesia.  As expected, due to Indonesia’s diverse topography and population, there are variations in the number of cases amongst its provinces.  Therefore clustering is needed to develop a map of COVID-19 cases to enable optimal handling of this pandemic. The provinces are clustered using K-means method according to their respective COVID-19 case numbers. Data taken from Indonesian Ministry of Health in November 2020 is used in this study, covering COVID-19 cases in Indonesia’s 34 provinces. K-means results in seven optimal clusters with variance ratio of 0.185. Clusters 1 to 3 cover most provinces in Java, including DKI Jakarta in Cluster 1 as the province with the most cases. Each of Clusters 4 and 5 consists of 5 provinces, while each of Clusters 6 and 7 consists of 10 provinces.  Cluster 7 comprises provinces with lowest cases of COVID-19.