Analisis Clustering Pelaku Usaha UMKM Kota Bekasi Menggunakan Algoritma K-Means

Abstract

Micro, Small, and Medium Enterprises (MSMEs) are the backbone of Indonesia's economy, employing over 97% of the workforce. However, the absence of data-driven classification hampers the formulation of effective support policies. This study aims to cluster MSMEs in Bekasi City based on business capital, revenue, and subdistrict location using the K-Means Clustering algorithm. A total of 7,923 entries were analyzed after cleaning the initial 7,964 dataset. The optimal number of clusters was determined using the Davies-Bouldin Index (DBI), with the best score of 0.5940 achieved at K=3. The clustering was conducted in two stages: manual calculation on a sample of 138 data points and automated processing via Python for the full dataset. The manual process converged at the 3rd iteration, where each step involved calculating Euclidean distances to cluster centroids, followed by centroid updates based on the mean of cluster members. The results classified MSMEs into three categories: small-scale, medium-scale, and large-scale enterprises. Small-scale MSMEs, typically with capital and revenue below IDR 5 million, are concentrated in Mustika Jaya, Pondok Gede, and Bantar Gebang. In contrast, large-scale MSMEs with higher financial figures are mostly found in Bekasi Selatan, Medan Satria, and Rawalumbu. This clustering model offers a practical foundation for designing more targeted, spatially informed, and adaptive MSME development policies.