Custom data analysis
Assumptions
Where do these measurements systematically disagree with the configured model?
Response linear in X, with slope and offset fitted from the measurements.
Independent variable: X / unspecified · Dependent variable: Y / unspecified
Reference scale is an assumed response noise floor, not measured uncertainty.
- The chosen baseline and unit labels are user supplied, not physically verified.
- Independent-variable measurements are treated as accurate; response errors are assumed approximately independent and comparable in scale.
- A linear calibration with a fitted offset is appropriate for the reference region.
- The positive response reference scale is an explicit user assumption, not measured uncertainty.
Adequacy depends on the configured model, reference scale, sampling and coverage.
A supported candidate identifies statistical inadequacy of the configured model. It does not identify a physical mechanism, universal threshold or safe operating range. Unknown units remain unspecified.
Read the methodology →Bring your measurements.
Choose columns, name the quantities, and review a supported model. Full measurements stay in this browser; optional AI shares limited context only when requested.
The example uses temperature in °C and response in V. Suggested assumed response scale: 0.05 V. Confirm these labels and scale yourself.
Optional AI · describe your experiment
AI explains. ModelScope's analysis engine decides.
Your full dataset stays in the browser. Optional AI setup sends your description and column headers; explanation sends the selected finding, summary statistics and caveats to Google Gemini. Free-tier inputs may be used to improve Google products. Avoid sensitive descriptions and labels.