All frameworks and cheat sheets

Checklist

The precision footer for reports built on automated tags

Checklist of six lines to add to any report built on automated tags, from what the flag identifies to when it was last re-measured.
Six lines that let a reader judge whether a tag count means anything

A six-line footer for any research report that relies on AI tags, sentiment labels or other automated flags. It records what the flag identifies, the base rate, measured precision and recall with sample sizes, the population they were measured on and the date of the last check. With it, readers can tell whether a count like 340 pricing complaints is close to the truth. Koji makes the two audits behind it quick because every tag links to its transcript.

Where this appears

What's in this checklist

  • What the flag claims to identify, with a written inclusion rule
  • Base rate from a random audit, with sample size and date
  • Measured precision from the flagged audit, with sample size
  • Measured recall from the unflagged audit, with sample size
  • The population it was measured on, named specifically
  • When it was last re-measured

Use this image

Free to use in your docs, decks and posts under CC BY 4.0, as long as you credit Koji with a link to this page. To embed it, paste this into your page:

html
<a href="https://www.koji.so/visuals/precision-footer-automated-tags-checklist"><img src="https://www.koji.so/visuals/precision-footer-automated-tags-checklist/precision-footer-automated-tags-checklist.png" alt="Checklist of six lines to add to any report built on automated tags, from what the flag identifies to when it was last re-measured." width="1200" height="630" /></a>
<p>Source: <a href="https://www.koji.so/visuals/precision-footer-automated-tags-checklist">The precision footer for reports built on automated tags</a>, Koji</p>