Skip to content

Latest commit

 

History

History
102 lines (69 loc) · 2.35 KB

File metadata and controls

102 lines (69 loc) · 2.35 KB

Summary card and narrative report guide

This document explains the two reporting layers added for cross-disciplinary teams.

Why they exist

In many projects, the bottleneck is not computing a metric. The bottleneck is explaining to another person:

  • what kind of time-series dataset this is
  • what the main structural issues are
  • why those issues matter for the project
  • what the team should do next

The summary card and the narrative report are built for that step.

Summary card

Use the summary card when someone wants a short, practical readout.

Typical readers:

  • clinicians
  • product managers
  • study coordinators
  • operations partners
  • collaborators outside the method team

What it emphasizes:

  • executive summary
  • top structure axes in plain language
  • main watchouts
  • analysis opportunities
  • recommended next actions

API:

profile.to_summary_card_markdown()
profile.to_summary_card_json()

CLI:

EchoTime data.npy --format summary-card

Narrative report

Use the narrative report when someone wants a full prose explanation.

Typical readers:

  • domain experts reviewing a dataset handoff
  • coauthors writing methods-light sections
  • lab members onboarding to a dataset
  • project teams deciding validation strategy

What it emphasizes:

  • what the dataset is in everyday language
  • what stands out structurally
  • why that matters for common tasks
  • what could go wrong if structure is ignored
  • practical next steps
  • reliability and interpretation guardrails

API:

profile.to_narrative_report()

CLI:

EchoTime data.npy --format narrative

How to use them together

A good pattern is:

  1. Run the full profile for technical inspection.
  2. Export a dataset card for reproducibility.
  3. Export a summary card for the broader team.
  4. Export a narrative report when decisions need explanation.

Good defaults by audience

General cross-disciplinary audience

Use the default audience.

Clinical audience

Use audience="clinical" when you want the surrounding workflow and examples to stay clinically framed.

Product or operations audience

Use audience="product" or audience="operations" if the dataset is closer to traffic, demand, or operational telemetry.

Neuroscience audience

Use audience="neuroscience" when you want the surrounding examples to match neural time-series work.