PDAM Performance Clustering using K-Means
Taufiqurrakhman Nur Hidayat, Fendi Aji Purnomo, Yudho Yudhanto · 2022
To increase the coverage and quality of drinking water services, PDAM conditions must be healthy to operate the Drinking Water Supply System (SPAM) effectively and efficiently through strong PDAM internal management. The evaluation of the performance of the PDAM's SPAM implementation annually produces PDAMs that have the performance of Healthy, Unhealthy, and Sick. The first thing to do is to determine the prioritized areas needed for further handling and improvement of PDAM performance. Clustering the provinces in Indonesia based on the categories of Healthy, Unhealthy, and Sick can solve this problem. The researcher proposes a K-Means model that uses the Square Euclidean Distance. The dataset used is PDAM Performance data sourced from the Agency for the Improvement of the Implementation of the Drinking Water Supply System (BPPSPAM) from the Ministry of Public Works and Public Housing website. The results show that the resulting model can classify PDAM performance between the features of Healthy with Unhealthy, Unhealthy with Sick, and Healthy with Sick. The classification looks more distributed in the relationship between the Unhealthy and Sick features. In cluster I, there are 14 data or provinces; in cluster II, there are ten provinces, and four provinces are in cluster III. So the total number of provinces included in the cluster is 25, and nine provinces are not included in any cluster.