Free browser tool · Browser-only processing

Scatter Plot Maker

Create a scatter plot from X-Y data pasted from Excel or a CSV. Add a line of best fit with Pearson r and R², then download PNG, SVG or CSV.

  • ✓True numeric X and Y axes
  • ✓Excel paste, CSV or TSV upload and column mapping
  • ✓Up to six color-coded groups
Your numbers are processed in browser JavaScript — nothing is uploaded.
⊹

Your scatter plot will appear here

Paste X–Y data on the left and click Generate scatter plot

— imported rows · — valid pairs plotted
Linear Association
—
Pairs (n)
—
Pearson r
—
R²
—
Association
Correlation is not causation. A scatter plot reveals patterns in the data you entered — it cannot establish that one variable causes the other. Confounding variables, reverse causation, or coincidence may explain any observed association.
DATA VISUALIZATION · GUIDE

Scatter Plots: What They Show and How to Read Them

Use the free scatter plot maker above to visualize any paired X–Y dataset. This guide explains what each result means, how to interpret correlation and regression, and when a scatter plot is the right chart to reach for.

What Is a Scatter Plot?

A scatter plot displays paired numerical observations as points on two axes. Each point has an X coordinate and a Y coordinate. You plot it because you want to see the relationship between two measurements taken from the same subject — not just what each one looks like on its own.

The pattern of dots tells you things that averages cannot: whether the values move together or in opposite directions, whether the relationship is roughly straight or curves, whether there are clusters suggesting distinct subgroups, and whether any individual points sit far from the others.

This free scatter plot maker takes X–Y pairs and renders them at their true numeric positions on both axes. Every calculation — correlation, regression slope, R² — runs in your browser. Nothing is uploaded to any server.

How to Use the Scatter Plot Maker

  1. Choose an input method. Paste two columns copied from Excel or Google Sheets, upload a CSV or TSV file, or load one of the built-in example datasets to get started quickly.
  2. Map your columns. After uploading a file, select which column is X and which is Y. You can also pick an optional grouping column to color-code points by category.
  3. Add chart labels. Give the chart a title and name both axes — including units helps anyone reading the chart understand what the numbers represent.
  4. Check the trendline option if you want to see the ordinary least-squares regression line and its equation. The scatter points always remain visible behind it.
  5. Click Generate scatter plot. The chart appears on the right alongside Pearson r, R², and a plain-language interpretation of what the numbers suggest.
  6. Adjust chart options to set manual axis bounds, change point size or opacity, or toggle gridlines. The chart updates immediately.
  7. Download. Export a PNG for presentations, an SVG for vector editing, or a CSV of the plotted data.
Input format: Paste tab-separated or comma-separated values with or without a header row. The tool detects both automatically. Enter 1234 not 1,234 — commas separate columns, not thousands. Negative values and decimals are supported. Rows with a missing X or missing Y are excluded and counted separately, not silently dropped.

Worked Example

Load the default positive-correlation example and you get this fictional dataset:

X: 1, 2, 3, 4, 5  |  Y: 2, 4, 5, 4, 5
5Pairs (n)
0.6Slope (b)
2.2Intercept (a)
≈ 0.7746Pearson r
0.6R²

The regression equation is Ŷ = 2.2 + 0.6X. At X = 1 the model predicts Y = 2.8; at X = 5 it predicts Y = 5.2. The actual Y values scatter around that line. R² = 0.6 means the fitted line accounts for 60% of the variation in observed Y values within this particular dataset — not a guarantee of future predictive power.

Pearson r ≈ 0.77 indicates a moderate positive linear association. As X increases, Y tends to increase too, though the relationship is not perfect.

How to Read a Scatter Plot

Positive association

Points drift upward from left to right. Higher X values tend to coincide with higher Y values. The classic real-world example is height and weight measured on the same people — taller individuals tend to weigh more, on average, though many exceptions exist.

Negative association

Points slope downward from left to right. Higher X values tend to coincide with lower Y values. Price and quantity demanded in economics often show this pattern: as price rises, fewer units sell.

No apparent linear association

The points look like scattered buckshot with no obvious slope. Pearson r will be near zero. That does not mean the variables are unrelated — it means there is no obvious linear relationship in the data you plotted.

Nonlinear relationships

Sometimes the dots trace a curve rather than a line. A U-shaped pattern — low Y at both ends with higher Y in the middle — gives a Pearson r near zero even though the relationship is strong. This is why looking at the scatter plot matters at least as much as reading the correlation number.

Load the "Nonlinear (U-shaped)" example in the tool to see this directly: r ≈ 0, yet the pattern is obvious on the chart.

Clusters

Groups of points that separate from each other suggest subpopulations in the data. Combining them into a single pooled regression can be misleading — the apparent overall trend may not describe any of the groups accurately.

Unusual observations

A point far from the bulk of the data deserves investigation. It could be a genuine extreme value, a measurement error, or a data-entry mistake. Check the original source before deciding what to do with it. Removing points to improve correlation is rarely a sound approach.

Adding a Line of Best Fit

The trendline this tool draws is an ordinary least-squares (OLS) regression line — the straight line that minimizes the sum of squared vertical distances from each point to the line. Check the "Show line of best fit" option before generating the chart.

The equation takes the form Ŷ = a + bX, where b is the slope (how much Y changes per one-unit increase in X) and a is the intercept (the predicted Y when X equals zero). Both the equation and R² appear below the chart when a trendline is enabled.

Regression slope
b = Σ(xᵢ−x̄)(yᵢ−ȳ) / Σ(xᵢ−x̄)²

Intercept: a = ȳ − b·x̄

The trendline spans the observed X range of your data. It is not automatically extended beyond that range, because extrapolating assumes the linear pattern continues in territory you have not measured.

Note: the regression line predicts Y from X. If you swap the columns, the slope and intercept will generally change. Pearson r itself is symmetric — it does not care which variable is on which axis.

What Pearson r Means

Pearson r measures linear association between two numerical variables. It runs from −1 to +1. A value of +1 means the points fall perfectly on a line with positive slope. A value of −1 means the same for a negative slope. Zero means no linear relationship exists in these data — not that the variables are independent in general.

r valueDirectionLinear strength
0.9 to 1.0PositiveVery strong
0.7 to 0.9PositiveStrong
0.5 to 0.7PositiveModerate
0.3 to 0.5PositiveWeak
0 to 0.3PositiveNegligible
0 to −0.3NegativeNegligible
−0.3 to −0.7NegativeWeak to moderate
−0.7 to −1.0NegativeStrong to very strong

These labels are conventional starting points, not universal rules. What counts as a meaningful r depends heavily on the field, sample size, and what question you are trying to answer.

If all X values are identical, or all Y values are identical, r is mathematically undefined. The tool displays "Undefined" in that case rather than showing a misleading zero. With only two observations that have distinct X values, r will be exactly ±1 by definition — which is a mathematical property, not evidence of a real-world relationship.

Pearson r also says nothing about causation. A scatter plot reveals patterns in your data. It cannot tell you which variable drives which, whether a third variable explains the pattern, or whether the relationship would hold in a different sample.

What R² Means

R² (the coefficient of determination) tells you what proportion of the variation in observed Y values is accounted for by the fitted regression line within this dataset. An R² of 0.60 means 60% of the variation in Y is explained by the linear relationship with X in these particular observations.

For simple OLS regression with an intercept, R² equals the square of Pearson r. The remaining fraction (1 − R²) represents variation the linear model does not account for — other influences on Y, measurement noise, or a relationship that is not linear in the first place.

R² is undefined when all Y values are identical (there is no variation to explain). The tool reports this explicitly rather than returning 1 or 0.

Scatter Plot vs Line Graph

FeatureScatter plotLine graph
Primary purposeRelationship between two numerical variablesTrend or change over an ordered sequence
X axisNumeric position, true scaleOften time or a categorical sequence
Points connectedNo — connecting is wrong for scatter dataYes — the line is the message
Repeated X valuesSupported and plotted normallyMay require aggregation or care
Best forHeight vs weight, price vs demand, test scoresMonthly sales, stock prices, temperature over time

The most common mistake people make with scatter data is connecting the dots. That is appropriate for a line graph where the order of observations matters. On a scatter plot, the dots represent paired measurements — connecting them in input order produces a meaningless zigzag. This tool never connects scatter points automatically.

If your X axis is time or a fixed sequence and you want connected points, the Line Graph Maker is designed for that purpose.

When to Use a Scatter Plot

Reach for a scatter plot when you have two numerical measurements from the same subjects and you want to explore whether and how they relate. Common uses across different fields:

FieldExample XExample Y
EducationHours studiedExam score
MarketingAd spendRevenue
ScienceTemperature (°C)Reaction rate
HealthcareAgeBlood pressure
EconomicsGDP per capitaLife expectancy
EngineeringLoad (kg)Deflection (mm)

A scatter plot is not the right choice when your X variable is categorical (use a bar chart or box plot), when you want to display how a single variable distributes (use a histogram), or when you want to show change across an ordered time series (use a line graph).

If you have more than two variables, scatter plots with grouping colors can encode a third dimension, but they cannot replace more specialized multivariate methods. See the guide on exploratory data analysis for how scatter plots fit into a wider analytical workflow.

Common Scatter Plot Mistakes

Treating unequal X values as equally spaced

If your X values are 1, 2, and 10, the gap between 2 and 10 should be eight times the gap between 1 and 2 on a properly scaled axis. Many simple chart tools treat these as three equally spaced categories, which distorts the visual relationship. This tool uses a true numeric linear scale.

Sorting X and Y independently

Each point represents one paired observation. Sorting the X column independently of the Y column breaks the pairs and produces a chart that shows a false relationship — or destroys a real one. Every X and Y value here stays paired to its original record.

Connecting scatter points

Connecting dots in input order creates a line that shows the sequence in which you entered the data, not any relationship between the variables. Scatter points should not be connected.

Calling r a percentage

Pearson r is not a percentage. r = 0.8 does not mean "80% correlation." R² is the proportion of variance explained, which is a different thing entirely.

Claiming correlation proves causation

This cannot be overstated. A scatter plot shows co-variation. It cannot identify the cause. Ice cream sales and drowning rates are positively correlated in many datasets — both rise in summer, for the same underlying reason (hot weather), not because one causes the other.

Removing outliers to improve correlation

Deleting inconvenient data points to make a pattern look stronger is not good practice. Investigate outliers — they may reveal something real about the data or its collection. Remove a value only if you have a specific, documented reason to believe it is erroneous.

The guide on data visualization mistakes covers related pitfalls across chart types. For a broader framing of where scatter plots sit in analytical work, see the data analysis methods overview.

Frequently Asked Questions

Can I paste data from Excel or Google Sheets?

Yes. Select two columns in Excel or Google Sheets, copy them, and paste into the Paste tab. The tool handles tab-separated and comma-separated values with or without a header row.

Can I upload a CSV or TSV file?

Yes. Use the Upload tab and drop or browse for a file up to 1 MB. After upload you can select which column maps to X, which to Y, and an optional grouping column.

Do scatter plot axes have to start at zero?

No. Forcing a scatter plot to start at zero often compresses the variation you are trying to see. The default auto-scaling uses appropriate bounds for the data. If you need zero included for context, use the "Include zero" options under Chart options.

Can X values repeat?

Yes. Multiple records with the same X are all plotted. If several points land at the same coordinates, a tooltip reveals how many observations share that position.

What happens if a row is missing X or Y?

Incomplete pairs are excluded from the chart and statistics. The count badge shows how many rows were excluded and why. The tool never shifts values to fill missing coordinates.

Can I show multiple groups?

Yes. After uploading a CSV with a category column, select it as the group field. Each group gets a distinct color. You can toggle group visibility using the series list.

Is the correlation coefficient the same as R²?

No. Pearson r measures linear association and ranges from −1 to +1. R² is r squared and ranges from 0 to 1. R² describes the proportion of Y variation explained by the regression line. For simple OLS, R² = r² exactly.

What does "Pearson r: Undefined" mean?

If all your X values are identical, or all your Y values are identical, the standard formula divides by zero. The correlation is genuinely undefined in that situation — not zero — so the tool says so explicitly.

Is my data uploaded anywhere?

No. All parsing, charting, and statistical calculations run in your browser. Nothing you enter or upload is sent to any server.

How is the trendline calculated?

Using ordinary least squares, which minimizes the sum of squared vertical residuals between observed Y values and the line. The slope and intercept are computed from valid paired observations only.

Can I download the scatter plot?

Yes — PNG for documents and presentations, SVG for vector editing, and a CSV of the plotted X–Y pairs.

What is the difference between a scatter plot and a bubble chart?

A bubble chart encodes a third numerical variable by sizing each point. This maker produces standard scatter plots with fixed-size markers. If you need bubble chart functionality, that is a separate chart type.

Does correlation imply causation?

No. A scatter plot reveals patterns in observed data. Establishing causation requires controlled experiments, careful study design, and domain knowledge — not a high r value.

When this tool is useful

Two variables. One plot. Clear answers.

This free scatter plot maker works best when you have two numerical measurements from the same subjects and want to see quickly whether they relate — and how strongly. It does not perform hypothesis testing, control for confounders, or prove statistical significance.

Relationships between two variablesTrendlines and regression equationsComparing groups on one chartExploring data before formal analysis
FAQ

Scatter Plot Maker questions

How the free browser tool works.

Is the Scatter Plot Maker free?+

Yes. It needs no account or signup for files within the supported size limit.

Can I paste data from Excel or Google Sheets?+

Yes. Copy two columns and paste them. Tab- and comma-separated values work with or without a header row.

Why are my X values not evenly spaced?+

The chart uses a true numeric scale, so the gap between 2 and 10 is eight times the gap between 1 and 2. Equal spacing would distort the relationship.

Is my data uploaded to a server?+

No. Parsing, charting, statistics and exports all run in your browser.