JevMade Sign in
← Back to experiments

Apps & data pipelines

jev-clinical-cleaner

Turn inconsistent research spreadsheets into standard columns, with code handling numbers and uncertain Jev suggestions flagged for human review.

Source screenshot of jev-clinical-cleaner
SOURCE SCREENSHOTFull screenshot ↗

What it does

The dry run uses a string-matching stand-in, not Jev. Its 0.9 review threshold is a setting to test on labelled data, not 90% accuracy or clinical validation.

How you can use it

Start with the made-up spreadsheet and example list of columns and allowed answers. Ask a developer to try the documented dry run first. It shows the cleaned table, items to check and a record of decisions without calling Jev. It matches words against your list instead, so it does not test Jev’s judgments.

For a research table, remove identifying details before any online run: column headings, sample values and cell text can leave your computer. Check both automatically accepted results and flagged items. Use examples people have already checked to decide where the tool should stop accepting answers automatically. Keep experts responsible for clinical meaning; this workflow is not evidence of diagnostic accuracy.

Keep this for later

Sign in to bookmark experiments, guides and videos, and keep notes only you can see.

Continue to sign in

We’ll bring you back to this listing.