Segmentation of West Java Province Based on Socio-Economic Indicators Using K-Means and Agglomerative Clustering Methods

Abstract

Regional segmentation based on socio-economic indicators is a crucial approach in data-driven development planning. Through accurate segmentation, governments can design more targeted policies aligned with the specific characteristics of each region. This study aims to compare two clustering methods, namely K-Means and Agglomerative Clustering, in grouping regions within West Java Province based on socio-economic indicators such as poverty rate, open unemployment rate, and Human Development Index (HDI). The analysis was conducted using the Python programming language on the Google Colab platform. Cluster performance was evaluated using the Elbow Method to determine the optimal number of clusters and the Silhouette Score to assess cluster quality. The results indicate that Agglomerative Clustering produces more consistent and interpretable segmentations, particularly in reflecting the socio-economic similarities between regions. However, in terms of computational efficiency, the K-Means method performs better due to its faster processing time. These findings offer valuable insights for regional policymakers in setting development priorities more effectively, grounded in the actual socio-economic conditions of each area