Jurnal Ilmiah Multidisiplin Indonesia
Vol. 5 No. 08 (2026): Jurnal Ilmiah Multidisplin Indonesia (JIM-ID), August 2026

Clustering the Sales Levels of Thrift Clothing on the Shopee Platform Using the K-Means Algorithm

Putra Dwi Kurniawan (Universitas Stikubank (Unisbank) Semarang)
Theresia Dwiati Wismarini (Universitas Stikubank (Unisbank) Semarang)



Article Info

Publish Date
25 Aug 2026

Abstract

The Shopee e-commerce platform has recorded exponential transaction growth, with Indonesia becoming its largest global market, while the thrift clothing segment has emerged as one of the significant sales categories within the platform. Thrift clothing sellers face challenges in objectively identifying daily sales performance patterns because transaction data are highly fluctuating. This study applies the K-Means algorithm within the Knowledge Discovery in Databases framework to cluster daily thrift clothing sales levels based on the variables of Total Sales and Total Orders. Primary data were obtained from the Shopee Seller Center of a thrift clothing store during the period from December 2025 to May 2026, resulting in 182 daily records, which were filtered into 168 active operational days after excluding 14 inactive days. Numerical variables were normalized using the Min-Max Normalization method before being processed with K-Means in RapidMiner Studio, while the optimal number of clusters was determined using the Elbow Method. Cluster quality evaluation was conducted using the Davies Bouldin Index in RapidMiner and the Silhouette Score, which was independently validated using Orange Data Mining. The results indicate that K = 3 is the optimal number of clusters, with a Davies Bouldin Index of 0.693 and a Silhouette Score of 0.512, both categorized as good. The clustering process produced three categories: Low Cluster (101 days, 60.1%, average IDR 134,139/day), Medium Cluster (50 days, 29.8%, average IDR 362,131/day), and High Cluster (17 days, 10.1%, average IDR 873,521/day). The High Cluster was concentrated in the April May 2026 period and contributed 31.9% of the total six-month revenue. These findings provide an empirical basis for thrift clothing sellers to formulate more targeted promotional strategies, inventory management, and marketing budget allocation based on historical data patterns.

Copyrights © 2026






Journal Info

Abbrev

esaprom

Publisher

Subject

Agriculture, Biological Sciences & Forestry Computer Science & IT Earth & Planetary Sciences Engineering Physics

Description

Jurnal Ilmiah Multidisiplin Indonesia (JIM-ID) is a peer-reviewed journal regularly published by the SEAN Institute every three months. namely, several research publications to publish multi-disciplinary articles with general topics on engineering, science, agriculture, plantations, forestry and ...