How do you handle missing or anomalous data in your marketing datasets before running an analysis for Square?
