<![CDATA[Abstrak - Pertumbuhan bisnis coffee shop yang pesat di Indonesia mendorong kebutuhan pengelolaan data transaksi secara cerdas guna mendukung pengambilan keputusan strategis. Penelitian ini mengimplementasikan algoritma Apriori dalam kerangka Market Basket Analysis untuk menganalisis pola kombinasi produk makanan dan minuman pada Element Street Coffee, Kodim, Semarang. Data yang digunakan berjumlah 4.790 baris yang tervalidasi menjadi 2.314 transaksi unik mencakup 40 jenis produk, terdiri dari 24 varian minuman dan 16 pilihan makanan, selama periode 30 September 2025 hingga 17 April 2026. Praproses data meliputi seleksi, normalisasi nama produk, transformasi ke format transaction basket, serta pemisahan 1.549 transaksi multi-item (66,9%) sebagai basis analisis. Parameter minimum support ditetapkan 2% (31 transaksi) dan minimum confidence 20% melalui pendekatan eksperimen iteratif. Penerapan algoritma menghasilkan 18 frequent 1-itemset, 21 frequent 2-itemset, dan 15 association rules yang terevaluasi. Kopsu Caramel (28,73%) dan Coffee Latte (28,08%) mendominasi sebagai produk paling sering dibeli. Meskipun kombinasi Coffee Latte dan Kopsu Caramel mencatat support tertinggi (7,23%), nilai lift di bawah 1 (0,8962) menunjukkan independensi statistik antar keduanya. Rules dengan nilai lift tertinggi ditemukan pada Creamy Latte mengarah ke Kopsu Caramel (lift=1,0211) dan Dimsum (4) mengarah ke Coffee Latte (lift=1,013), mengindikasikan preferensi pembelian lintas kategori yang genuine. Temuan penelitian ini direkomendasikan sebagai dasar strategi bundling menu, cross-selling, penempatan produk, dan efisiensi manajemen stok bagi manajemen Element Street Coffee. Kata kunci : Algoritma Apriori; Association Rules; Data Mining; Market Basket Analysis; Pola Kombinasi Produk; Abstract - The rapid growth of the coffee shop business in Indonesia drives the need for intelligent transaction data management to support strategic decision-making. This study implements the Apriori algorithm within a Market Basket Analysis framework to analyze food and beverage product combination patterns at Element Street Coffee, Kodim, Semarang. The dataset consists of 4,790 rows validated into 2,314 unique transactions covering 40 product types—24 beverage variants and 16 food options—during the period of September 30, 2025 to April 17, 2026. Data preprocessing included selection, product name normalization, transformation into transaction basket format, and isolation of 1,549 multi-item transactions (66.9%) as the analytical basis. A minimum support of 2% (31 transactions) and minimum confidence of 20% were established through iterative experimentation. Application of the algorithm yielded 18 frequent 1-itemsets, 21 frequent 2-itemsets, and 15 evaluable association rules. Kopsu Caramel (28.73%) and Coffee Latte (28.08%) dominated as the most frequently purchased products. Although the Coffee Latte and Kopsu Caramel combination recorded the highest support (7.23%), a lift value below 1 (0.8962) indicates statistical independence between the two. Rules with the highest lift values were found in Creamy Latte to Kopsu Caramel (lift=1.0211) and Dimsum (4) to Coffee Latte (lift=1.013), indicating genuine cross-category purchasing preferences. These findings are recommended as the basis for menu bundling strategies, cross-selling, product placement, and inventory management efficiency for Element Street Coffee management. Keywords: Apriori Algorithm; Association Rules; Data Mining; Market Basket Analysis; Product Combination Pattern;]]>
