Last updated
A hockey data warehouse for clubs, without a data team
Short answer
A club warehouse stacks raw provider loads, a shared identity layer (one player ID) and analysis-ready marts in a dedicated EU database the club owns. MotherDuck or DuckDB in an EU region is a practical default; Power BI, Excel, Sheets, Streamlit and Claude read from the same marts.
Raw → identity → marts
Raw tables preserve provider schemas. Identity resolves players, games and teams across sources. Marts are what sporting directors and analysts actually join — one row per player with league, tracking and scouting fields aligned.
Why DuckDB / MotherDuck
Many clubs do not want to operate Snowflake day one. DuckDB-family engines in EU regions keep cost and ops low while still giving Power BI and Python/Streamlit a stable SQL endpoint.
Destinations
No connector is live yet. The Hockey Brain lists every named product as Planned or Exploring — intended compatibility only, not a partnership.
- Power BI and Tableau (Planned)
- Excel and Google Sheets (Planned)
- Streamlit or Python notebooks on the same database
- Claude or ChatGPT via MCP (Planned)
Frequently asked questions
Do we need a data engineer on staff?
You need someone who can own access and licences on the club side. The Hockey Brain maintains pipelines during the founding program so that person is not rebuilding exports weekly.
Where is the warehouse hosted?
In the EU, in an account the club controls. Processing and matching also run on EU infrastructure.
Can we use Snowflake instead?
Yes. Snowflake and BigQuery are Planned delivery targets — same marts, different destination.
Get a free map of your club’s data stack
30 minutes · hello@thehockeybrain.com