K-Means Analysis for Clustering Regencies/Cities in West Java Based on Poverty and Divorce as Indicators of Welfare

Abstract

Community welfare is an essential indicator in assessing the success of regional development, not only through economic aspects such as poverty levels but also through social indicators such as divorce, which reflects family stability and the psychosocial condition of society. This study aims to map the welfare conditions of regencies/cities in West Java Province using a quantitative approach with the K-Means Clustering method and four main variables: the number of divorce cases due to disputes, divorce due to economic factors, domestic violence (KDRT) cases, and the percentage of poor population based on data from the Central Bureau of Statistics (BPS) in 2024. All data were analyzed through stages of data cleaning, normalization, and clustering, with the selection of three optimal clusters using the elbow method to ensure representativeness and avoid bias. The results indicate a pattern of regional grouping that reflects different welfare characteristics. Areas in Cluster 0 tend to have high divorce rates due to disputes but low poverty levels, illustrating non-economic social pressures. Cluster 1 is dominated by regions with high poverty levels and divorce cases caused by economic factors, indicating significant financial vulnerability. Meanwhile, Cluster 2 consists of regions with relatively high domestic violence indicators, suggesting stronger domestic pressures compared to other factors. These findings emphasize that welfare is influenced not only by economic variables but also closely related to family dynamics. Therefore, K-Means-based clustering can serve as a strategic reference in formulating more targeted public policies in West Java Province.