Apriori Algorithm Integration in a Laravel-Based POS System for SME Purchase Pattern Analysis

Purpose: This study develops and evaluates a Laravel-based Point of Sale (POS) application integrated with the Apriori algorithm to uncover customer purchase patterns and generate actionable association rules for Sehatea, a beverage micro-enterprise operating in Bandar Lampung, Indonesia.

Research Methodology: A system development research design was applied through four sequential phases of planning, design, coding, and testing. A total of 131 validated transaction records were digitized and analyzed using the Apriori algorithm with a minimum support threshold of 20% and a minimum confidence threshold of 20%, and system functionality was verified through black-box testing across six test cases.

Results: Eight association rules were generated. The strongest rules show that customers who purchase Iced Taro Tea also tend to purchase Iced Green Tea (confidence 37.50%, support 2.29%), while Plain Water purchasers frequently co-purchase Iced Tea Original (confidence 36.36%, support 3.05%). Iced Tea Original dominated single-item frequency with 55.73% support, and all six system functions passed testing.

Conclusions: The integrated system delivers real-time, data-driven insight that supports bundling, cross-selling, and inventory decisions.

Limitation: Findings are constrained by the single-site, small-sample context.

Contributions: The study offers a replicable, low-cost architecture for embedding data mining into SME point-of-sale systems in developing economies.