Data Point Evidence (Trustworthy Data)

A model is only as good as its data. Real measurements sometimes go wrong: a defective sample, an interrupted reading, a miscalibrated instrument. SenSight lets you attach an evidence record to any row:

A status flag: Normal / Suspect / Invalid
A reason + free comment
A photo (from your phone camera)
A reference-sample (CRM) check with auto-computed deviation

Rows flagged Invalid are automatically excluded from regression — and the exclusion is documented in your exported reports, so the analysis stays defensible in an audit.

Try it in the app

Try: In the Data tab, click the ✎ next to a row → set it "Invalid" with a reason → run regression. Note the banner showing N rows excluded, and check the "Data Point Notes" section in an HTML/Excel export.

Data Point Evidence (Trustworthy Data)

Run this on your own data

Pareto Analysis (The 80/20 Rule)

Gage R&R & MSA (Can You Trust Your Measurements?)