SQL Playground
I wanted to answer my own product questions in the data without waiting on anyone, so I built a realistic SaaS database to learn on, and made it open source.
An open-source SQL playground: a realistic SaaS product, modelled and seeded so you can query it like the real thing. I built the whole data model and dataset in a day, to get fluent in the questions that actually drive product decisions.
Dashboards are great for the metrics you already track. The interesting questions are the ones you have not thought to ask yet, and answering those means querying the raw data yourself. I did not want to keep waiting on data-team capacity every time I had one.
Most SQL courses teach syntax against toy tables that look nothing like a real product. I wanted to learn on data shaped like the thing I actually work on: a subscription SaaS, with all the messiness that brings.
So I built one. SQL Playground is an open-source PostgreSQL database, hosted on Supabase, modelled as a realistic subscription SaaS:
- Five connected tables: users, plans, subscriptions, invoices and events
- Seed data with the messiness of a real product: trials, upgrades, churn and refunds
- A set of exercises and worked example queries, so anyone can pick it up and learn the same way
The whole thing is open source on GitHub: schema, seed data, exercises and example queries, so anyone learning product analytics can clone it and start querying.
Working through it myself surfaced exactly the kind of insight it was built for: users who hit a key activation milestone in their first week were far more likely to stick around and upgrade.
I built it solo in about a day. I leaned on AI to explain errors and suggest cleaner query syntax, but kept the analytical thinking, the data modelling and the question-asking, firmly mine. The point was not to generate queries, it was to get fluent in reasoning with data.
Turning raw product data into decisions is exactly the foundation a lot of businesses are missing. I can stand up the data model, wire up the pipeline and, crucially, ask the questions that move the metric, not just build another dashboard.
Want something like this built for your business?
I turn real problems into working products, fast, with AI. Let’s talk.
Start a conversationStart a conversation