Practice joins, aggregation, window functions, and data modeling with downloadable datasets. Compare table structures and choose a practical project.
Use ecommerce data to work through joins and sales reporting. Choose subscriptions for time-based analysis, or inventory for reconciliation across operational records.
| Dataset | Best suited to | Structure | SQL setup | Main challenge |
|---|---|---|---|---|
| Ecommerce Sales & Returns | Joins, aggregation, and sales reporting | 6 related tables | Tested SQLite file | Preventing duplicated revenue across order lines and refunds |
| SaaS Subscriptions & Churn | Retention, cohorts, and window functions | 5 related tables | Tested SQLite file | Defining active status from one subscription period per account |
| Marketing Campaigns & Conversions | Attribution and mixed-grain analysis | 4 related tables | Tested SQLite file | Joining daily spend to individual leads without fanout |
| Inventory & Supply Chain | Reconciliation and operational modeling | 7 related tables | Tested SQLite file | Explaining balances from opening stock and signed movements |
| Bike Sharing Demand | Time-series aggregation and seasonal demand | 2 related tables | Tested SQLite file | Comparing hourly and daily grains correctly |
Run this against the Bike Sharing SQLite database. It keeps the hourly grain and compares observed hours and average rentals by working-day status.
SELECT
workingday,
COUNT(*) AS observed_hours,
ROUND(AVG(cnt), 1) AS average_hourly_rentals
FROM hour
GROUP BY workingday
ORDER BY workingday;Write down the result you want before joining anything. Identify what one output row should represent, which tables contain the required information, and whether each relationship can create multiple matches. Check the row count and totals as you build the query.
Begin with a filtered selection or grouped result. Add a related table when the question requires it. As the analysis grows, separate repeated logic into named steps and document the definitions that determine the answer.
No. Download the published SQLite database and open it in a compatible local SQL tool. Start with the available files rather than building a larger platform just to complete the first exercise.
Continue with a free SQL exercise that helps you check your joins, grouping, or time-based logic.