Build tested SQL operations analytics case study

[mjeans]

Closes #1

What changed

• added a normalized SQLite schema and deterministic synthetic database generator
• created reusable member-activity and site-month metric views
• added data-quality, activation, cohort-retention, site-performance, support, and anomaly queries
• saved all aggregate query outputs for immediate review
• added a decision memo that translates results into operational next steps without causal overclaiming
• added five integration tests and GitHub Actions validation

Why

This case study demonstrates practical SQL analysis from data model through business communication, including joins, CTEs, conditional aggregation, date logic, cohort analysis, and window functions.

Local validation

• rebuilt 8 sites, 480 members, 3,076 service events, and 252 support tickets from scratch
• regenerated six query outputs
• all data-quality checks returned zero failures
• all five integration tests passed
• workflow YAML, Python syntax, and CSV structures validated

All records are deterministic and synthetic.

[mjeans]

Validation complete: the GitHub Actions workflow rebuilt the deterministic demo database, ran all saved SQL queries, and passed the five automated tests. Ready to merge.