Outlier: Track previous outliers when constructing aggregate dataframe

[brifordwylie]

Right now the logic will just take top 10s, put them together, and the drop_duplicates(). The issue with this is that you miss the opportunity to pull different outliers.