Production healthcare data

API
DOPPEL turns governed DataHub datasets into privacy-safe synthetic twins.

This production dataset contains sensitive information.

DOPPEL creates a useful development twin without exposing the source records.

Production / Restricted Clinical Operations · Patient Data Team
Seeded demo scale 1.00× · seed 42 · expiry 30 days · publish on

Generation plan

Before execution, DOPPEL reads DataHub context and chooses a strategy for every field.

Seed42
Scale1.00×
Expiry30 days
PublishDataHub

Live pipeline

Connecting to backend…

  1. 01
    Read DataHub contextSchema, tags, ownership, relationships
  2. 02
    Build generation planStrategy selected per governed field
  3. 03
    Generate patientsIdentifiers replaced, distributions retained
  4. 04
    Generate encountersRelationship-preserving foreign keys
  5. 05
    Check privacyDirect identifiers, exact-row leakage, singling-out
  6. 06
    Check utilityDistributions, correlations, aggregates
  7. 07
    Check relationshipsDiagnosis/procedure/claim associations
  8. 08
    Check referential integrityForeign-key validity, cardinality
  9. 09
    Publish to DataHubLineage, tags, scores, evidence
Result —

Waiting for pipeline result.

Exact copied rows—source rows reproduced unchanged
Direct identifier matches—names, emails, IDs, phones, addresses
Privacy score—direct-identifier and row checks
Utility score—distribution and schema fidelity
Referential integrity—foreign-key validity
Singling-out risk—quasi-identifier uniqueness
Correlation similarity—pairwise numeric correlations
Conditional similarity—grouped conditional means

Want to see what a failing run looks like?

DataHub writeback

The synthetic twin is registered in DataHub with lineage and evidence.

Tags
SYNTHETIC NON_PRODUCTION DOPPEL_VERIFIED
Evidence receipt

Run evidence is linked to each synthetic asset via InstitutionalMemory.

Expiry

—

Owner / domain

—

View in DataHub Download twin Download evidence