🏠 Dashboard · ☑ Progress · 🔎 Finder · 📚 Block · ← Previous · Next →
Sequence 238 — Data Quality deepening - break, root-cause, repair and prevention
Context: Chapter 13 - Data Quality Management | Applied Lab Deepening
Source: CDMP Master Execution Manual
When / purpose
Use controlled defects to distinguish remediation from prevention and connect quality work to governance accountability.
What you should learn
- Repairing a row is not the same as fixing the process.
- Root cause requires evidence.
- Preventive checks target the source/process failure.
- Before/after evidence matters.
What to read / study
- MUST: Chapter 13 applied lab folder — use the break/repair cycle; do not overwrite raw data.
- REFERENCE: SQL/Python technology evidence — use working copies/staging and version-controlled scripts; no production ETL.
Do this in order
- Choose 2–3 documented defects.
- Trace likely root cause using source/system/process evidence.
- Repair a controlled working copy.
- Add a preventive validation/control.
- Re-run the profile and compare results.
- Identify owner/steward/escalation implications.
Meridian application
Connect the quality result to Meridian governance domains/data products if the platform work has begun.
Create / save
Before/after profile + root-cause note + remediation/prevention evidence.
Stop / boundary
Stay inside the current source/skill boundary; no production ETL or engineering depth.
Completion gate
You can explain why the preventive control addresses the failure mechanism rather than only the symptom.
NEXT → Sequence 239