Analysis of Protein Consumption Data and Desired Dietary Patterns as a Basis for Provincial-Level Food Security Information Systems Using the K-Means Algorithm

Abstract

Food security and nutritional status, particularly in efforts to eliminate stunting, are important issues in Indonesia. Stunting caused by chronic malnutrition is greatly influenced by low protein consumption, especially during the first 1,000 days of life. This study aims to analyze the relationship between average per capita protein consumption and the Food Consumption Pattern Score (FCPS) at the provincial level from 2021 to 2023, as well as to explore the role of information systems in supporting food security policies. Data were obtained from data.go.id and analyzed using descriptive statistics, Spearman's correlation, and K-Means clustering methods. Results showed a significant positive correlation between protein consumption and PPH scores (ρ = 0.604; p < 0.001), indicating that protein intake is closely related to dietary diversity. Cluster analysis yielded two main groups: a low cluster dominated by eastern Indonesia, and a high cluster including Yogyakarta and Jakarta. Although the national PPH score increased from 81.81 (2021) to 84.96 (2023), inter-regional disparities remain high. These findings underscore the need for cluster-based interventions and the use of information systems to support more informed decision-making. The limitations of the data, which are not yet fully curated, highlight the need for further studies considering socio-economic variables.