Pemetaan Pola Produktivitas Tanaman Pangan Menggunakan Data Warehouse dan K-Means Clustering

Saputra, Fernando Adi
Santoso, Cahyono Budy

TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 6 No 10 (2026): March 2026

The productivity of food crops between regions shows significant differences despite having large areas of planted and harvested areas. These inequalities reflect production patterns that are not yet well structured, which makes it difficult to analyze and make data-driven decisions. This research aims to evaluate the productivity of food crops using K-Means, based on planted area and harvest area and productivity, with the aim of overcoming irregular production patterns between regions and producing grouping patterns that are more systematic and easy to analyze. The method used is KDD (Knowledge Discovery in Database). Data obtained from the official https://jabar.bps.go.id/id and https://opendata.jabarprov.go.id/, website for the 2019-2022 period. The application of data warehouses with star schema yields a Tabel of productivity facts, and dimensions of years, crops and regions. Metric evaluation using the Davies-Bouldin Index showed good cluster quality information. This research resulted in the application of data warehouse, clustering using the K-Means algorithm with a Davies-Bouldin Index score of 0.419 with optimal k = 3, the amount of data in cluster 0 is 368 data, cluster 1 is 52 data, and cluster 2 is 12 data. To simplify interpretation and support data-driven decision making, business intelligence tools are used to display clustering results. The results of this research help provide strategic recommendations in increasing agricultural productivity of food crops in West Java through region-specific interventions.