<![CDATA[Deforestation remains one of the major environmental challenges in Indonesia because the rate of tree cover loss varies considerably among regions, making it difficult for stakeholders to identify priority areas for forest monitoring and management using conventional descriptive analysis. This study aims to identify spatial patterns of deforestation by clustering Indonesian districts/cities based on multi-year Tree Cover Loss and to evaluate the effectiveness of the K-Means clustering algorithm for supporting data-driven environmental analysis. The dataset was obtained from Global Forest Watch (GFW) and consists of Tree Cover Loss data for 402 districts/cities in Indonesia during 2023–2025, represented by three numerical attributes measured in hectares. The research methodology includes data cleaning, attribute selection, Min-Max normalization, determination of the optimal number of clusters using the Elbow Method, K-Means clustering, and cluster evaluation using the Davies–Bouldin Index (DBI). Experimental results show that the optimal number of clusters is three, producing 333 districts (82.84%) in the low-loss cluster, 61 districts (15.17%) in the moderate-loss cluster, and 8 districts (1.99%) in the high-loss cluster, with a DBI value of 0.6599, indicating good clustering quality. The findings reveal that tree cover loss is unevenly distributed across Indonesia and provide a data-driven regional categorization that can support priority setting for forest monitoring and conservation. The scientific contribution of this study lies in utilizing multi-year Tree Cover Loss data (2023–2025) at the district/city level across Indonesia to characterize regional deforestation patterns using K-Means clustering.]]>
