When encountering a column similar to 'FBI CD' from the Crimes dataset, the current implementation treats it as numeric and right now it's flagging almost the whole column as mismatch. Should it be treated as numeric or categorical? This is another question which the implementation is unclear on.