2026-05-21 Better Sample Data Meeting notes

Anonymization scripts status

• Context: Would like to complete work on anonymization scripts, run them on an MSU dataset as a means of demonstrating successful PPI removal for MSU’s Risk Management office

• Lee needs volunteer developer time to complete the scripts

• Update project call to include use of AI as a possible tool

• Will post call on Slack channels

Generating a dataset instead of seeking a contributed/anonymized one

• Given the difficulty we’ve had making headway on a contributed dataset, is generating random data the better approach?

• For users and bib data, generating random user profiles would not be too challenging

• It’s historical data that is the problem, and the record of cross-app interactions

  • I.e., loan rules and check-outs related to instances

  • Fiscal year budgets and ledgers associated with certain orders

• Could AI be a tool for generating the more complex data and data interactions, if fed the appropriate JSON schema info?

  • Shelley would like to investigate, as time permits

Presentation from GBV/Index Data colleagues in June

• A look at their approaches to populating reference environments with good sample data

• MODINVUP-184: GBV. MIU. Add sample data to be populated in the reference environmentsClosedPreview