Abstract
The development of new pipe connections is one of the important aspects in improving clean water services provided by the Regional Water Company (PDAM). However, the determination of priority areas for pipe network development at PDAM Tirta Silaupiasa Asahan Regency is still carried out subjectively, so it is less objective and data-based. Therefore, this research aims to apply the K-Medoids clustering algorithm to group development areas of new pipe connections based on their characteristics. The data used consisted of 22 sub-district areas with three input variables, namely the number of prospective new customers, distance to the existing pipe network, and population density. The data was normalized using the Min-Max method, then processed using the K-Medoids algorithm with the number of clusters k = 3 and Euclidean Distance as the distance measure. The clustering results were evaluated using the Silhouette Score method. Based on the test results, the clustering process converged after three iterations with the total cost decreasing from 6.3947 to 5.3436. The clustering produced three groups: Cluster 1 (High Priority) with 1 area, Cluster 2 (Medium Priority) with 16 areas, and Cluster 3 (Low Priority) with 5 areas. The average Silhouette Score obtained was 0.3981, which is included in the reasonable structure category. The system was implemented as a web-based application using PHP and MySQL, so it can help PDAM in determining priority areas for pipe network development objectively and based on data.
