<![CDATA[Transaction data at automotive workshops and accessory stores holds valuable information about customer purchasing habits, yet is typically used only as an administrative archive without further analysis to support business strategy. This is the case for Impala Variasi, an automotive workshop in Welahan, Jepara, which recorded 1,388 transactions from January 2025 to May 2026 but lacks a mechanism for identifying products customers tend to choose together, leaving service packages based on the owner's intuition. This study applies the FP-Growth algorithm to discover association patterns among components attached to each service, following the CRISP-DM framework and evaluating results using support, confidence, and lift ratio, with a minimum-support threshold of 5% (70 of 1,388 transactions). Results, cross-verified against raw data, show 13 frequent items, 7 frequent 2-itemsets, and 1 frequent 3-itemset (21 itemsets total), yielding 14 association rules. Speaker → Head Unit produced the highest confidence among the genuinely bidirectional rules (81.12%, support 8.36%, lift 7.51), followed by Head Unit → Speaker (77.33%) and Car Shampoo → Wax Polish (75.31%, lift 8.57). Three pairs — Oil Filter/Engine Oil, Spark Plug/Engine Oil, and Tire Valve/4-Tire Nitrogen — showed one-directional 100% confidence, meaning every customer who bought the smaller add-on also bought the base item, while the reverse direction (e.g., Engine Oil → Oil Filter, 53.42%) reveals meaningful upsell headroom. These patterns indicate consistent attach behavior between core services and add-ons, plus a strong bidirectional link between car-audio components. Results can guide service bundles, cross-selling, sparepart stock, and data-driven recommendations for Impala Variasi's management.]]>
