Introduction If you work with large, diverse data sets, you know that there will invariably be noise and erroneous data in the database you are using.  This is one reason why I treat all data as guilty until proven innocent.