<![CDATA[Stunting remains a complex public health problem, thus requiring a spatial approach to identify patterns of case distribution more precisely. This study aims to analyze the spatial pattern of stunting cases in Piyungan Subdistrict through a local hotspot approach, global spatial autocorrelation, and individual point-based micro-clustering. This study used spatial data and electronic medical records analyzed using ArcGIS software. Local hotspot identification was conducted using the Getis-Ord Gi* method, global spatial autocorrelation was analyzed using Moran’s I, while micro-clustering was analyzed using Average Nearest Neighbor (ANN). The results of the Getis-Ord Gi* analysis of 132 spatial units showed that 131 units (99.2%) were not classified as statistically significant hotspots or coldspots, while 1 unit (0.8%) was identified as a local hotspot at the 95% confidence level with a z-score of 2.21 and a p-value of 0.0267. Moran’s I analysis produced a value of -0.0365 with a z-score of -0.6623 and a p-value of 0.5078, indicating the absence of significant global spatial autocorrelation so that the distribution pattern of stunting cases at the aggregate level tended to be random. However, the ANN analysis showed an observed mean distance of 83.88 meters, an expected mean distance of 236.57 meters, a nearest neighbor ratio of 0.3545, a z-score of -20.214, and a p-value of <0.001, indicating the presence of very strong spatial clustering at the micro level. These findings indicate that the spatial pattern of stunting in Piyungan Subdistrict depends on the scale of analysis; at the administrative level, no strong cluster was found, whereas at the individual level, there was significant case clustering. Thus, the results of this study confirm the importance of more targeted nutritional interventions in micro-clusters to improve the effectiveness of stunting prevention and management.]]>
