This document analyzes consumer transaction data using association rule mining to understand purchasing patterns. It pre-processes the sparse dataset by pruning items with less than 2% support. Association rules are generated at different support and confidence levels, with more rules found at lower thresholds. The top rules show related purchases. A decision tree also predicts dairy purchases, with some common rules between the unsupervised and supervised models. Association mining is recommended for market basket analysis due to its ability to handle sparse data and generate simple, interpretable rules for cross-selling opportunities.