First Notebook in 5 Minutes

1. Create a project

From the Workspace home, create a new project. Every notebook, file, and execution artifact lives inside a project-scoped workspace.

2. Create a notebook

Add a new notebook to the project and open it. Each notebook runs in its own isolated session — installed packages and variables don’t leak between notebooks or between runs.

3. Write your first cell

Every code cell has access to the sm module without any imports:

df = sm.read_csv("sample.csv")
df.head()

sm.path("relative/path") resolves any path safely inside your project workspace — use it any time you read or write a file directly instead of through sm.read_csv / sm.write_csv.

4. Run it

Run the cell (or the whole notebook). Output — dataframes, plots, printed text — renders inline. Every run is traced: SAMSKARA keeps the code, output, and any files the cell wrote as an execution artifact you can revisit later.

5. Write a result

sm.write_csv(df, "output/summary.csv")

Files land under the project’s workspace and are visible in the file browser immediately after the cell finishes.

Next steps

• Connect a Database to pull in real data instead of uploaded files
• ArrowLake to write results into a real Iceberg table instead of a flat file
• Scheduler to run this notebook on a cron schedule