CLUSTERING OF DISTRICTS/CITIES IN INDONESIA BASED ON POVERTY LEVEL USING THE K-MEANS METHOD WITH THE ELBOW METHOD APPROACH (CASE STUDY: BPS 2025 DATA)

Abstract

Poverty remains a significant structural issue in Indonesia's development. The disparity in poverty levels across regions highlights the need for a data-driven analytical approach to comprehensively understand its distribution patterns. This study aims to cluster districts and cities in Indonesia based on the number of poor residents using the K-Means Clustering method optimized via the Elbow Method. The data used consists of secondary data from the Central Statistics Agency (BPS) for the year 2025, covering 514 districts/cities across 38 provinces. The analysis process began with a data preprocessing stage, including data cleaning, outlier detection using the Interquartile Range (IQR) method, and normalization using StandardScaler. The optimal number of clusters was determined using the Elbow Method and evaluated using the Silhouette Score. The analysis results show that the optimal number of clusters is K = 3 with a Silhouette Score of 0.6987, which falls into the "good" category. The classification resulted in three groups: Low Poverty (369 regions or 71.8%), Moderate Poverty (113 regions or 22.0%), and High Poverty (32 regions or 6.2%). Although the number of regions in the High Poverty group is relatively small, this group accounts for 26.4% of the total national poor population and is dominated by regions with high population density on the island of Java. These findings are expected to serve as a basis for the government in formulating more targeted and data-driven poverty alleviation policies