Cluster Analysis of Districts and Cities in North Sumatra Province Based on Socio-Economic Indicators

Abstract

The success of development program implementation related to community welfare in North Sumatra Province can be evaluated through several indicators, including consumption patterns, housing and environment, population, education, nutrition, health, expenditure, and employment. This study aims to group districts and cities in North Sumatra Province based on similar or homogeneous characteristics of these welfare indicators. To achieve this, both hierarchical and non-hierarchical clustering methodswere applied, including k-means and k-medoids (using PAMand CLARA algorithms). Among these methods, the hierarchical clustering using Ward’s linkage was found to produce the best clustering results, forming three distinct clusters. The first cluster consists of 4 districts/cities, the
second cluster includes 23 districts/cities, and the third cluster comprises 6 districts/cities. These
clustering results can assist local governments in North Sumatra in formulating more targeted and
effective welfare policies, based on the dominant indicators present in each cluster.