ACTUAL VS PREDICTED DIAGNOSTIC
Residual Plot Generator
Paste observed values and the predictions that go with them. This tool plots each residual (observed − predicted) against its supplied predicted value and shows an accessible row table. It does not generate predictions itself.
YOUR DATA
Enter observed and predicted values
Paste or upload two numeric columns: observed values and the predictions supplied by a model or another calculator. Use a period as the decimal mark. Up to 5 MB and 5,000 observations are accepted.
The file is read locally in your browser.
YOUR RESULTS
Residual plot
Residuals versus supplied predicted values
The horizontal dashed line marks zero residual. Points are the pairs you supplied.
Residuals by observation
Residual = observed − predicted.
| Row | Observed | Predicted | Residual | Absolute residual |
|---|
How to use this plot
- Paste paired observed and predicted values, or load the example.
- Choose which column is observed and which is predicted.
- Inspect the residual plot for systematic patterns such as curvature, fan-shaped spread or outliers.
- Download the residual table as JSON or CSV for further analysis.
What the plot shows
For each pair the residual is observed − predicted. The horizontal axis shows the supplied predicted value; the vertical axis shows the residual. A random cloud around the zero line is consistent with no obvious pattern in these residuals, while a trend or changing spread suggests the model may be missing structure or variance heterogeneity.
RMSE and MAE are shown to summarise residual magnitude. They are computed from the pairs you supplied; this page does not fit a model or generate predictions.
Limitations
- Predictions are supplied by you, not generated by this tool. The plot can only reveal patterns in the values you upload.
- Observed and predicted values must be aligned row by row and use the same units.
- A pattern in the residuals is a diagnostic clue, not proof that a particular assumption is violated.
- The current input limit is 5,000 observations and 5 MB per file.
Related tools
RMSE calculator · R-squared calculator · Compare regression models