🔒 Privacy Lab

Anonymize a CSV and adversarially test the result - powered by the data-anonymization-toolkit.

Workflow:

Upload your CSV (or use the bundled sample)
Edit the YAML config (column lists, k-anonymity threshold, noise)
Run anonymization
Switch to the Red Team tab and run 10 adversarial attacks against the output

The toolkit drops direct identifiers, generalises quasi-identifiers, injects calibrated noise, scrubs fingerprints, and enforces k-anonymity. The red team then probes for residual re-identification, linkage, and fingerprinting risks.

Upload CSV (leave empty to use bundled sample)

Sample CSV

Anonymization config (YAML)

First 50 rows

Download full anonymized CSV

Legend: 🔴 CRITICAL | 🟠 HIGH | 🟡 MEDIUM | 🟢 LOW

The 10 attacks: Uniqueness, Temporal Linkage, Fingerprints, Outlier Re-identification, Distribution Skew, Null Pattern Linkage, Rare Combo Linkage, Numeric Precision, Numeric Ratio, Compound Entity.