Skip to main content
Every analysis produces a workspace inside SkoutLab. Think of it as the analyst’s notebook: the SQL, charts, notebooks, and exports generated while testing hypotheses. This page shows how to use it.

Accessing the Workspace

From any report:
1

Open a finding

Click a hypothesis card to open the drawer.
2

View Evidence

Scroll past the narrative to see charts, tables, and the “Evidence Package” button.
3

Download

Choose Download Workspace to grab everything for that hypothesis, or Download Evidence Package to get a zip of the curated artifacts.

What’s Inside an Evidence Package

ItemDescription
notebooks/analysis.pyGenerated Python notebook or script used during validation.
charts/*.pngHigh-resolution charts referenced in the report.
summary.mdMarkdown summary of the finding and statistics.
data-samples.csvAnonymized sample rows supporting the hypothesis.
Packages are zipped so you can archive them, attach to tickets, or send to auditors.

Live Workspace Files

You can also browse workspace files without downloading:
  • Charts API/api/charts/... streams chart images so the UI loads quickly.
  • Workspace Files API/api/workspace-files/... returns the list of scripts/notebooks generated.
  • Evidence Package API/api/evidence-package/... bundles everything server-side before downloading.
All routes require auth and automatically include your X-User-ID, so only the owner of an analysis can fetch its workspace.

When to Use It

Audit & Compliance

Keep the evidence zip as proof of how numbers were produced.

Handoff to Data Team

Share the notebook/SQL so engineers can productionize a finding.

Manual Deep-Dive

Open the raw workspace if you want to rerun or tweak the analysis locally.

Tips

  • Each hypothesis has its own workspace. Download only what you need.
  • The zip includes metadata (analysis ID, hypothesis ID) so files are traceable.
  • Workspace files are kept for the lifetime of the analysis. Deleting the report removes the workspace immediately.
Want these analyses to run on a cadence? Head to Automation & Scheduled Analyses →