Anonymization of profiling results via Edge

As profiling results are stored in Collibra, anonymization can be an important tool for your company's security requirements. The anonymization happens at the end of the profiling process via Edge, before the results are sent to Collibra.
Anonymization means that values are replaced by random hash values. Identical values in a column get the same hash value so that you can still recognize the values as identical.

Collibra provides two types of anonymization of profiling results via Edge:

• By default, profiling results are automatically anonymized for columns with the Text or Geo data type.
  In this case, we anonymize data distribution charts, Mode attributes, and Percentiles.

  Note Edge detects the data type of a column during profiling and anonymizes the results only if the data type attribute is Text or Geo. However, if Edge detects a data type that doesn't correspond with the actual data type, data may not be anonymized or may have been wrongfully anonymized. You can update the data type in the Metadata tab of the Descriptive Statistics section on the Summary tab of a Column asset.
  

• An administrator can decide to anonymize the profiling results for all columns. This can make sure that all sensitive data is anonymized.
  In that case, we:
  • Anonymize frequency charts, categories, and Mode attributes.
  • Don't store or show data distribution charts.
  • Don't store or show Percentiles.
  • Don't store or show basic statistics such as Mean, Median, Variance, Standard deviation, Minimum value, and Maximum value.

  

  The required setting to activate this is "Anonymize Edge profiling results for all data types".