urn:li:dataset:(urn:li:dataPlatform:kafka,payments.authorized_transaction_events,PROD)
→
urn:li:dataset:(urn:li:dataPlatform:postgres,fraud.realtime_payment_features,PROD)
01
Proposed change
- Source column
device_trust_score- Scenario
device-trust-range-change- Source asset
urn:li:dataset:(urn:li:dataPlatform:kafka,payments.authorized_transaction_events,PROD)- Description
- The device-risk producer will change device_trust_score from a 0–100 range to a 0–1 range without updating downstream expectations.
02
DataHub impact path
Discovered from recorded DataHub MCP context; the request did not name a model.
urn:li:dataset:(urn:li:dataPlatform:postgres,fraud.realtime_payment_features,PROD)
→
urn:li:mlFeature:(realtime_payment_risk_features,device_trust_score)
urn:li:mlFeature:(realtime_payment_risk_features,device_trust_score)
→
urn:li:mlModel:(urn:li:dataPlatform:mlflow,payment_fraud_detector_v3,PROD)
Affected production model
urn:li:mlModel:(urn:li:dataPlatform:mlflow,payment_fraud_detector_v3,PROD)payment_fraud_detector_v3
03
Bounded experiment plan
Reproduce the proposed device-score range change across the full evaluation population and measure its effect on the downstream fraud model.
| Operator | Column | Fraction | Value |
|---|---|---|---|
scale_numeric | device_trust_score | 100% | 0.01 |
04
Baseline vs mutated metrics
Experiment 1 · scale_numeric
7.3 msMutation applied to 500 of 500 evaluation rows.
05
Deterministic rule evidence
Numeric policy alone makes the decision.
auc_drop: 0.059783 > 0.050000
prediction_flip_rate: 0.974000 > 0.150000
06
DataHub write-back
Published to DataHubRecorded remote status: Succeeded
- Result tag
Provenance-BLOCK- DataHub document
urn:li:document:shared-e04fccf3-0be4-4cde-aadb-5a1237daa4e2
Recorded during a live local DataHub run. This hosted page presents committed evidence and does not contact or mutate DataHub.