A Hybrid Approach Using K-Means and Apriori Algorithms for Marketing Recommendation Systems in the Café Industry

Abstract

Many food and beverage (F&B) businesses struggle to survive because of a lack of understanding of their market and customer behavior. Therefore, this study aims to analyze transaction data using K-Means Clustering and Apriori algorithms to identify customer purchasing patterns and provide actionable insights for a café. This research applies the CRISP-DM method, including six primary phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. K-Means reveals that products can be grouped into three main clusters, namely the low-demand products, premium products, and best-selling products. Furthermore, the association rule mining successfully indicates six meaningful correlations among products, with the strongest lift value of 5.22 was found between Piccolo Latte and Extra Oatmilk. These findings are transformed into actionable insights, such as recommendations for implementing bundling and up-selling. An interview with the business owner validates that the recommendations are considered applicable to the current business operations. This proves that the integration of K-Means and Apriori algorithms can be used to design effective data-driven marketing strategies for an F&B business.