Data governance
Quality checks, reconciliation and stewardship
Deterministic data-quality rules run against every load, with source-to-target reconciliation proving that what left the source is what landed in the warehouse. The rule catalogue and results below are fabricated for this portfolio lab.
Quality score
75%
Checks failing
2 / 8
Reconciled domains
3 / 4
Classified domains
4
Data-quality check catalogue
Synthetic results| Check | Dimension | Dataset | Rows | Failed | Rate / threshold | State |
|---|---|---|---|---|---|---|
| DQ-01Billing ledger — mandatory account key present | completeness | FCT_BILLING_LEDGER | 1,284,500 | 0 | 0% / 0% | pass |
| DQ-02Customer master — natural key uniqueness | uniqueness | DIM_CUSTOMER | 412,880 | 3 | 0.0007% / 0.01% | warn |
| DQ-03Currency code conforms to ISO-4217 reference set | validity | FCT_BILLING_LEDGER | 1,284,500 | 214 | 0.0167% / 0.05% | warn |
| DQ-04Telemetry rollup totals reconcile to raw partitions | consistency | AGG_DEVICE_TELEMETRY | 8,940,000 | 9,812 | 0.1098% / 0.05% | fail |
| DQ-05SCD2 dimension has exactly one current row per key | consistency | DIM_CUSTOMER | 412,880 | 0 | 0% / 0% | pass |
| DQ-06Landing files arrive inside the ingest window | timeliness | STG_BILLING_FEED | 96 | 4 | 4.1667% / 2% | fail |
| DQ-07Restricted attributes tokenised before persistence | validity | DIM_CUSTOMER | 412,880 | 0 | 0% / 0% | pass |
| DQ-08Archive candidates carry a checksum proof | completeness | ARC_LEDGER_HISTORY | 640,200 | 128 | 0.02% / 0.5% | warn |
Source-to-target reconciliation
Billing
reconciledSynthetic billing extract → FCT_BILLING_LEDGER · 2026-08-12
Source rows1,284,500
Target rows1,284,500
Row variance (tolerance)0 (0)
Amount variance0.00
Customer
reconciledSynthetic CRM snapshot → DIM_CUSTOMER · 2026-08-12
Source rows412,883
Target rows412,880
Row variance (tolerance)-3 (5)
Amount variance0.00
Telemetry
varianceSynthetic device gateway → AGG_DEVICE_TELEMETRY · 2026-08-12
Source rows8,940,000
Target rows8,930,188
Row variance (tolerance)-9812 (4000)
Amount variance0.00
Archive
reconciledLedger history staging → ARC_LEDGER_HISTORY · 2026-08-12
Source rows640,200
Target rows640,200
Row variance (tolerance)0 (0)
Amount variance0.00
Domain stewardship and classification
| Domain | Steward | Classification | Retention | Masked in non-prod |
|---|---|---|---|---|
| Billing transactionsAmounts retained; counterparty identifiers tokenised outside finance roles. | Finance data steward (role, not a person) | confidential | 7 years, then provable purge | Yes |
| Customer masterTax and contact attributes tokenised in-flight; only tokens are persisted. | Customer domain steward (role) | restricted | Active + 24 months after closure | Yes |
| Device telemetryDevice identifiers hashed; no subscriber linkage held in the rollup layer. | Platform steward (role) | internal | 13 months rolling | Not required |
| Reference & conformance dataChange-controlled; every version keeps an effective-dated history row. | Data governance council (role) | public | Indefinite, versioned | Not required |