Context & Masking for Discovery Classification
The Quick Review modal helps users readily identify data samples across one or several data sources while comparing them against enabled likelihoods. The features available in Global Settings allows users to expand search context to better validate results and avoid unnecessary access to personal information while ensuring reviews are not interrupted.
Things to Know
-
Both context collection and masking are performed on the worker node. Therefore, these settings must be enabled on and the data source must be rescanned.
-
Only new scans would reflect the context and masking features. Data samples in existing scans will remain as they were.
To enable context and matching settings
- In the Classification Settings screen, enable your preferences for the masking and context features. If the data sample setting is enabled, the masking is OFF and context setting will be ON by default.
Important:
Changing settings requires a new scan to show
Setting
Description & Details
Turn on Data Masking
Data masking obscures sensitive info (N for numbers, A for letters). Only the masked sample is collected from data sources.
Special characters are displayed as they are.
Example: cbradshaw93@email.com would be AAAAAAAAANNN@AAAAA.AAA, which still helps the user validate that the result is an email address.
Include context near sample
Context is collected 20 characters before and 10 characters after the term.
Context collection can potentially expose sensitive information if the classified data has adjacent data.
To use settings during review
- On the Data Discovery menu, select . The Terms screen will present the total count of discovered terms. You can use the Select Data Source(s) field and the search bar to filter your results.
- Any masked data will appear with a highlight replacing any data with 'A' for alphabetical numbers and 'N" for numeric characters.