Data Validate Skill

Validate database readiness before annual report generation or data operations.

When to Use

• •Before generating an annual report
• •After running data migrations or imports
• •When troubleshooting missing or broken data
• •As a pre-flight check for any data pipeline

Workflow

1. Run Database Validation

cd annual-reports/scripts
python3 validate_data.py --year [YEAR]

This checks:

• •Required tables exist and have data (profiles, stories, board_members, partners)
• •Annual report eligible stories (need >= 5)
• •Financial data for current and previous year
• •Staff statistics for past 3 years
• •Leadership messages exist

2. Interpret Results

The script outputs color-coded results. For each failure:

• •Note the specific table/field that's missing
• •Check if the data exists but under a different column name
• •Check if a migration needs to be run

3. Self-Correcting Loop

If validation fails:

1. •Missing tables: Check supabase/migrations/ for unapplied migrations
2. •Empty tables: Check if data exists in a different schema or needs seeding
3. •Missing stories: Query stories table to understand what's available:
   sql

   SELECT count(*), annual_report_eligible, status
   FROM stories
   GROUP BY annual_report_eligible, status;
   

4. •Missing financials: Check annual_financials table structure matches expected schema
5. •Fix the issue using Supabase MCP tools
6. •Re-run validation to confirm the fix

4. Report

Output a summary:

• •Total checks passed/failed
• •Data completeness percentage
• •Specific gaps that need manual attention (e.g., leadership messages need human input)
• •Recommended next steps

Data Quality Checks

Beyond table existence, verify:

• •Stories have content (not just titles)
• •Stories have associated media (images)
• •Financial numbers are reasonable (no negative revenue, expenses match categories)
• •Board member names are properly formatted
• •Partner list is current
• •Staff statistics show reasonable year-over-year trends

Integration with Report Generation

After validation passes, the report pipeline is:

validate_data.py → assemble_content.py → generate_pdf.py → validate_pdf.py

Each step should pass before proceeding to the next.