<![CDATA[This study investigates the application of the K-Means clustering algorithm to classify Posyandu service areas based on toddler demographic characteristics, with the goal of supporting more efficient planning and targeted distribution of nutritional aid. Using a dataset consisting of 855 toddler records from the Puskesmas Braja Caka region, data preprocessing steps—including one-hot encoding, handling of categorical locality attributes, and Z-score standardization—were performed to ensure consistent feature representation. The Elbow Method indicated that six clusters provided the optimal balance between compactness and interpretability. The resulting cluster distribution comprised 133 toddlers in Cluster 0, 75 in Cluster 1, 151 in Cluster 2, 238 in Cluster 3, 125 in Cluster 4, and 133 in Cluster 5. Further analysis revealed distinct demographic characteristics: Clusters 0 and 2 had higher median ages, Cluster 3 displayed the widest age variability, and Cluster 4 showed the highest proportion of male toddlers. PCA visualization confirmed a clear separation among clusters, while boxplots illustrated meaningful differences in age distribution. These findings demonstrate that K-Means clustering effectively uncovers demographic patterns that can guide policymakers in allocating resources more accurately and prioritizing interventions for communities with higher toddler density or greater nutritional risk. As an actionable recommendation, health authorities are advised to prioritize nutritional supplementation and intensified monitoring in Cluster 3 (highest density, 238 toddlers) and Cluster 4 (male-dominant, youngest age group), while deploying tailored growth-monitoring programs in Clusters 0 and 2 where older toddlers are concentrated. This approach strengthens data-driven decision-making for Posyandu operations.]]>
