How would you handle missing data or outliers in a large dataset before feeding it into a predictive model?
