Regional Cluster Analysis Based on Community Consumption Patterns and Quality of Life for Market Segmentation Strategies Using the K-Means Algorithm

Abstract

This study addresses the challenge of regional inequality in Indonesia, where economic growth and consumption patterns do not always correspond to improvements in quality of life. Differences in food expenditure, non-food expenditure, and life expectancy across provinces create the need for more objective regional mapping to support market segmentation strategies. This study aims to classify 38 provinces in Indonesia based on community consumption patterns and quality of life using the K-Means clustering algorithm. The research employed a quantitative data mining approach using secondary data from the Central Statistics Agency in 2025, including life expectancy, average food expenditure, and average non-food expenditure. Data were cleaned, integrated, and transformed using Min-Max Scaling before clustering. The optimal number of clusters was determined using the Elbow Method and Silhouette Score, resulting in three regional clusters. The model produced an Average Silhouette Score of 0.5820 and a Davies-Bouldin Index of 0.6095, indicating valid cluster separation. The results identify three market categories: Basic Priority Areas, High Expenditure Areas, and Stable Growth Regions. Each cluster reflects different purchasing power, welfare conditions, and strategic market potential. The study concludes that K-Means clustering can provide a useful decision-support framework for businesses and policymakers in designing targeted, efficient, and region-based market segmentation strategies.