Analisis Pola Penjualan Toko Ritel NisaMart Menggunakan Algoritma K-Means Klasterisasi

Abstract

This study applies the K-Means algorithm to analyze product sales patterns in a retail store. The dataset consists of 100 transactions recorded on August 5, 2025, within a single operational day. Preprocessing was performed by converting transaction time attributes into categories of morning, afternoon, and evening. Clustering using the WEKA software produced three product groups based on sales levels. Cluster 0 contains products with the highest sales, averaging 6.67 units, and relatively low prices. Cluster 1 includes products with medium sales, averaging 4.45 units, and moderate prices. Cluster 2 consists of products with the lowest sales, averaging 2.54 units, but with the highest prices. The results indicate that the K-Means method is effective in identifying product segmentation, serving as a reference for inventory management and targeted marketing strategies. Future research is recommended to extend the data collection period and include additional variables for more representative analysis