Use this free residual plot error analyzer in your browser. No sign-up, no download and no upload: everything runs on your own device, so your data stays private.
What this tool does
A residual plot shows what a model got wrong, and the pattern of the errors matters more than their size. Random scatter around zero means a decent fit; a curve means a missing term; a funnel means the error grows with the value. This tool plots residuals from observed and predicted pairs and reports RMSE and mean residual, so the verdict is quantitative as well as visual.
How to use it
- Paste or type your observed-predicted pairs in the box above.
- Adjust any settings, such as cutoffs, scales or styles, to match your data.
- Click the run button and read the output below the tool.
- Check the summary numbers before you copy anything into a document.
- Copy the output, code or image into your figure, slide or report.
Worked example
Enter observed-predicted pairs 10/11, 20/19, 30/32, 40/38 and 50/51. The residuals are −1, 1, −2, 2 and −1, scattered around zero with an RMSE near 1.5, which reads as a good small-error fit. Replace the last pair with 50/40 and the final residual jumps to 10, a lone point far from the line.
Tips for better results
Plot residuals against predictions, not against time, unless time is your suspected problem.
Honest limits
The tool plots residuals you supply; it does not fit a model or estimate predictions. Your predictions must come from your own analysis.
Choosing the right settings
Start with the default values. They are chosen to show a sensible result at once, so you can see the layout before you touch your own data. Then change one setting at a time. Changing one thing at a time is the fastest way to learn what each control does, and it stops two changes from cancelling each other out. If the output looks wrong, check the input format first: a missing comma, a swapped column or a stray space causes most errors, and the tool reports what it received in its summary line.
Using the result in a figure
For a figure, keep the output simple. One message per figure, short labels, and a caption that states the method and the settings you used will serve readers better than a dense, decorated chart. If a journal or supervisor will see the result, note the tool name, the input values and any cutoff in your methods or caption, so the figure can be reproduced later. That habit also helps you: six months from now, the caption is the only record of how the figure was made.
Common questions
Do I need an account? No. The residual plot error analyzer runs entirely in your browser with no sign-up and no saved data.
Is my data uploaded anywhere? No. Your observed-predicted pairs stays on your device, which is why the tool is safe for unpublished work.
Can I use the output in a paper or presentation? Yes. The output is yours to reuse; just check it visually once, note the settings in your caption, and credit the method, not the tool, in your references.
What if the result looks wrong? Check the input format first, then change one setting at a time. Most errors come from a missing comma, a swapped column, or a value outside the expected range, and the summary line under the tool shows what it actually received.
Try the residual plot error analyzer above with the sample data, then with your own. It is free, it needs no account, and it works in any modern browser.
When you reuse the result, keep a short record of the inputs and settings you used. A one-line note in your project file, such as the data pasted, the cutoff chosen and the date, is enough to reproduce the figure later. If a collaborator questions a number, that note lets you rebuild the exact output in seconds instead of guessing what changed.
Related tools
Versión en español · All Free Online Tools · quantile quantile alignment visual.
Embed this tool on your website — free
Copy the code below to add this working tool to your own page. A small credit link stays with it — that’s all we ask.