Data Warehouse product
Lab
Python on tables you already cataloged
Run pandas, matplotlib, and plotly on bronze and silver tables, then save the result as a new silver dataset — explore before you model, without leaving the catalog.
Start with a CSV or a database. No data science team required.
Interactive notebooks
Python on your Lake tables
Load bronze or silver data, visualize with matplotlib or plotly, and persist transforms back to the catalog — all in the browser.
df = lake.read("bronze", "sales_q4")
df.head()monthly = df.groupby("region")["revenue"].sum()
monthly.plot(kind="bar")lake.save_silver("regional_revenue", monthly)Query Lake directly
Reference bronze and silver tables by name — no export step or local copies.
Charts inline
matplotlib, seaborn, and plotly render in the notebook so you see results immediately.
Save without overwriting
Persist notebook outputs as new silver datasets. Source tables stay untouched.
Capabilities
Explore before you model
Run Python on bronze and silver tables, visualize results, and save outputs back to your Lake catalog.
Python cells
Run pandas code against your bronze and silver Lake tables in a lightweight notebook.
Visualize
Render matplotlib, seaborn, and plotly charts inline without leaving Heimdall.
Save to silver
Persist notebook results as new silver datasets — existing tables are never overwritten.
Isolated execution
Cells run in sandboxed workers with read-only access to your data until you explicitly save.
Use cases
What teams build with Lab
Analysts and data engineers use Lab to profile, transform, and validate Lake data before gold publish.
Browse all warehouse and intelligence scenarios on the use cases hub.
Explore before you train
Analyze in Lab, curate in Lake, and publish gold datasets for ML, Forecast, and Loop.