Penentuan Menu Potensial Menggunakan Algoritma K-Means Pada Titiksetara Coffee Bekasi

This study aims to identify potential customers and help local coffee entrepreneurs compete by offering products that customers desire, as well as address challenges in determining high‑potential menu items using the K‑Means algorithm. The data consist of TitikSetara Coffee’s revenue records from October 2023 through January 2024. RapidMiner software was used to perform clustering via the K‑Means algorithm. The sales data were grouped into three clusters: Cluster 0 for high‑potential menu items, Cluster 1 for medium‑potential menu items, and Cluster 2 for low‑potential menu items. Out of the 26 records analyzed, Cluster 0 contains 4 menu items, Cluster 1 contains 14 menu items, and Cluster 2 contains 8 menu items, with a Davies–Bouldin Index of 0.374. The identified high‑potential menu items can serve as a reference for TitikSetara Coffee, and it is expected that these recommendations will be well‑targeted.