Data Quality Assessment

Data Profile Summary

Phase 2: Quality Dimension Assessment

Completeness (required fields populated)

Accuracy (values reflect reality)

Consistency (data matches across systems)

Timeliness (data available when needed)

Uniqueness (no unintended duplicates)

Validity (values conform to rules)

Phase 3: Issue Prioritization

1. Quantify business impact per issue
   • Revenue impact (incorrect pricing, missed orders)
   • Operational impact (failed processes, manual workarounds)
   • Compliance risk (regulatory data requirements)
   • Analytics impact (incorrect insights, poor ML models)
   • Customer impact (wrong communications, data errors)
2. Prioritize by impact and fixability

Issue Priority Matrix

Phase 4: Root Cause Analysis

1. Investigate top quality issues
   • Source system data entry errors
   • ETL transformation bugs
   • Missing validation rules at ingestion
   • Schema evolution without migration
   • Integration failures dropping or corrupting data
   • Stale reference data
2. Document root cause per issue

Phase 5: Remediation Plan

1. Fix existing data quality issues
2. Implement preventive controls
   • Input validation at source
   • Schema enforcement in pipelines
   • Data contracts between producers and consumers
   • Automated quality checks in ETL/ELT pipelines
   • Referential integrity enforcement
3. Set up data quality monitoring

Phase 6: Automated Quality Monitoring

1. Implement continuous quality checks
   • Automated quality tests in data pipelines (dbt tests, Great Expectations)
   • Quality score dashboards per domain
   • Anomaly detection for quality metrics
   • Alerting when quality drops below threshold
   • Quality trend reporting (weekly/monthly)
2. Define quality SLAs per domain

Counter-Rationalizations

Output Format

• Data Profile Report: Statistical profile of all assessed datasets
• Quality Scorecard: Score per dimension per dataset
• Issue Inventory: All issues with severity and root cause
• Remediation Plan: Fixes and preventive controls with timelines
• Monitoring Configuration: Automated quality checks and alerts

Action Items

• Run data profiling on all in-scope datasets
• Assess quality across all six dimensions
• Prioritize issues by business impact
• Investigate and document root causes
• Implement data fixes and preventive controls
• Deploy automated quality monitoring
• Establish monthly quality review cadence