If you were handed a messy dataset with thousands of SKU transactions, how would you begin your analysis to find pricing leakage?
