Scatterize 3

Navigate:
Plot settings

Comma-separated numbers, e.g. 0, 50, 100

Open a dataset…

or

Drop a CSV anywhere on the page · Data never leaves your browser

Examples: Datasaurus Dozen · ’74 Cars · Star Cluster CYG · Ames Housing

New to Scatterize? Read the introduction.

Diagnostics

Scatterize reference

Statistical models

OLS
Standard linear regression. Fits a line minimizing the sum of squared errors. Sensitive to outliers.
Robust
MM-estimation with Tukey biweight weighting. Automatically downweights outliers and resists high-leverage points that would distort an ordinary fit.
Spearman
Rank-based correlation. Measures monotonic association without assuming linearity or normality.
Theil-Sen
Median-based slope estimator. Highly robust to outliers; uses the median of all pairwise slopes.

Smoothers

Running median
Local medians in a sliding window. Robust to outliers; reveals trends without assuming a shape.
LOWESS
Locally weighted regression. A smooth curve fitted independently at each point.

Censoring

Click any point to censor it — exclude it from the regression while keeping it visible as a grey hollow marker. Click again to restore it. Use censoring to explore how individual observations affect your results.

Labeling points

Right-click (or shift-click) a point to give it a name that follows it as you change variables or models — handy for tracking whether an outlier is unusual everywhere or just on one axis. Keyboard users can navigate to a point and press e, or use the Label button in the navigate bar. Clear a label by emptying its text. Labels are saved in the URL and included in SVG, PNG, and CSV exports.

Privacy & sharing

Everything runs in your browser — your data is never sent to a server. If you load a public CSV or Google Sheet URL, all settings (variables, model, censored points) are encoded in the URL. Share that URL and anyone with access to the data will see the same view.

Press Esc to close

Keyboard shortcuts

Model & plot

u / iPrevious / next X variable
j / kPrevious / next Y variable
o / pPrevious / next model
l / ;Previous / next legend group

Navigate points

/ Previous / next point
/ Back / forward 1 percentile
/ First / last point
n / mChange sort criterion (X / Y / Q-Q)

Censor & label

eAdd / remove label from selected point
SpaceCensor / uncensor selected point
cClear all censored points

Press ? to open or close this help

Export CSV

Export SVG

SVG uses League Spartan — install the font to render text correctly.

Export PNG

Share URL