Analisa Cluster Data Transaksi Penjualan Minimarket Selama Pandemi Covid-19 dengan Algoritma K-means

Abstract

Covid-19 has had a negative impact on the economic sector in Indonesia. This can be seen from the losses experienced by industry players in the form of a decrease in income turnover. Sales strategy needs to be done so that losses can be minimized. Sales transaction analysis can be done to find product groups with the most sales data so that stock management can be fulfilled and increase sales transactions. Berkah Abadi Minimarket is an industry that has been affected by this pandemic. Data analysis has not been carried out to find out which product has the most sales data, so it is necessary to analyze it with the k-means algorithm. This algorithm can group data based on similar characteristics. The application of the algorithm on 278480 transaction data, obtained three sales data clusters, namely cluster 2 or the highest sales of 57 products, cluster 1 or moderate sales of 57 products and the rest are cluster 0 with low sales. The result of the accuracy of the clustering model generated by the confusion matrix is 87%. Based on these results, the owners of the Berkah Abadi Minimarket can be helped in making decisions on stock management while the Covid-19 pandemic is still ongoing.