Grouping Sales Levels Smartphone Of Offline Store Using BIRCH Clustering Algorithm

Abstract

From 2020 to 2024, TM_Store and Jaya Com exhibited different sales patterns based on cluster analysis using the BIRCH algorithm. The background of this research is to provide strategic insights to both stores for improving their sales performance through data analysis. The sales data used includes brand, type, month, year, stock quantity, quantity sold, unit price, and total sales. The BIRCH method was chosen for its effectiveness in handling large datasets and providing accurate clustering results. The clustering results indicate a significant increase in the "Moderate" category, from 12 sales in 2020 to 354 sales in 2023. Meanwhile, the "Very High" category also saw an increase from 5 sales in 2020 to 97 sales in 2023, with sales in the "Very Low" category remaining high at 70 sales in 2023. On the other hand, Jaya Com was dominated by the "Very High" category, with a sharp increase from 25 sales in 2020 to 597 sales in 2023. The "High" category also showed significant growth, from 6 sales in 2020 to 98 sales in 2023. This data indicates that Jaya Com focuses on high-performance products, while TM_Store shows a more balanced distribution across various sales categories. Based on the analysis, Jaya Com had 1988 data points with 1984 cluster points, whereas TM_Store had 2012 data points with 1811 cluster points. Overall, the study concludes that the BIRCH algorithm can identify significant sales patterns in both stores, aiding in the development of more effective and efficient promotional strategies tailored to each sales category's performance.