Multidimensional Cluster Analysis of the 2023 Child Protection Index (CPI) Based on Provinces in Indonesia

Abstract

This study aims to (1) classify Indonesia's 34 provinces into homogeneous clusters based on the multidimensional profile of the 2023 IPA, (2) map the structural relationships between IPA dimensions through network analysis, and (3) analyze the unique characteristics of each cluster to formulate focused and differentiated policy recommendations. This study uses a quantitative approach with the K-Means non-hierarchical cluster analysis method and is enriched with correlation network analysis. The data analyzed are the scores of the five dimensions that comprise the 2023 IPA released by the Ministry of Women's Empowerment and Child Protection (KemenPPPA) and the BPS-Statistics Indonesia. Determining the optimal number of clusters using the Elbow Method and Bayesian Information Criterion (BIC) shows three clusters are the most appropriate structure. The analysis results identified three clusters of provinces with significantly different characteristics: Cluster 1 ("Multidimensional Structural Challenges"), Cluster 2 ("Moderate and Even Performance"), and Cluster 3 ("Excellent Implementers"). Network analysis revealed powerful and positive relationships between the Family Environment dimension (D2) and Special Protection (D5) and Basic Health (D3), as well as between Civil Rights (D1) and Basic Health (D3). This result indicates that the family environment and access to health care are central foundations supporting various other aspects of child protection. Combining these two analyses provides a more nuanced understanding than a single national ranking