Implementasi Algoritma Apriori Untuk Analisis Pola Pembelian Konsumen Pada Umkm Bestea Berbasis Website

Abstract

Digital transformation has encouraged micro, small, and medium enterprises (MSMEs) to adopt information technology in managing business activities, particularly in transactions and data analysis. umkm Bestea faces challenges in understanding consumer purchasing patterns, which are essential for marketing strategies and inventory management. Manual analysis is often ineffective due to limited resources and data complexity. Therefore, data mining, especially the Apriori algorithm, is applied as a solution to discover product associations frequently purchased together, thereby improving operational efficiency and supporting data-driven decision-making. This research employed the Extreme Programming (XP) software development method, which consists of planning, design, coding, and testing phases. The dataset comprises 2,129 sales transactions recorded by umkm Bestea between September and December 2024. The Apriori algorithm was used to calculate support and confidence for each itemset to identify frequent itemsets that meet minimum thresholds. A web-based system was developed using the Laravel framework, featuring purchase pattern analysis, product management, stock, composition, and reporting modules. The system was evaluated using black-box testing to ensure that all functionalities worked properly. The results show that the Apriori algorithm successfully identified consumer purchasing patterns, such as the tendency of Iced Green Tea being purchased together with Iced Tea Original or Iced Lemon Tea. The developed system not only supports purchase pattern analysis but also functions as an integrated Point of Sale (POS) application. This research contributes to assisting MSMEs in formulating promotional strategies, optimizing inventory, and enhancing the effectiveness of data-driven decision-making.