Hi David,
Thanks for the questions.
As you pointed out, working with anonymized data can be very challenging especially when the matching rules contain fuzzy comparison instructions. One of the approaches that I have used is to create sample data that can be shared with the team for testing match rules from the raw data by simply shuffling individual records' data elements (attributes) such as First Name, Last Name, Address, Date Of Birth, etc. Additional checks can be put in place to ensure that after shuffling the data elements, any record that looks the same or similar to the raw records is not selected in the sample data set. For match rule with exact comparison, we should continue to use the simple string replacement to hide the sensitive PII data.
Hope this helps. Please let me know if you have any follow-up questions on this.
------------------------------
Suchen Chodankar
------------------------------
Original Message:
Sent: 04-14-2022 09:57
From: david cessot
Subject: PII anonymization in lower environment and impact on match rule testing
Hello Community,
I would like to know how you succeeded in managing both privacy needs (by anonymizing PII information in testing environment of a source system) and preserving match rule testing capability. If PII is anonymized in lower environment, then you usually cannot test match rule unless some specific actions are done like:
- consistent anonymization (using dictionaries like replacing all 'John' by 'Mike' for example), but then you might loose fuzzy match rules
- encrypting in the lower environment source system, and having some decrypting capability on Reltio side
Thank you
------------------------------
david cessot
Schneider Electric
Grenoble
------------------------------