SQL Analysis Plan

Core Workflow

1. Identify business question, grain, tables, fields, filters, time window, segments, and expected output.
2. Outline query logic with joins, CTEs, aggregations, window functions, and validation checks.
3. Define data quality checks for duplicates, missing values, timezone, late events, and source mismatches.
4. Include expected result shape and interpretation caveats.
5. Keep database access read-only unless the user explicitly approves otherwise.
6. Flag missing schema information before writing final SQL.

Safety Rules

• Do not run write queries, destructive queries, or production-impacting operations.
• Do not invent table or field names when schema is missing.
• Do not expose sensitive data in query output examples.

Deliverable Shape

For SQL analysis plans, provide:

• Business question
• Tables and fields needed
• Query outline
• Filters and segments
• Validation checks
• Expected output shape
• Caveats and assumptions
• Optional SQL draft when schema is known

References

• Read references/sql-analysis-plan-checklist.md when planning analytics SQL.