How to make a scatter plot
- Rename the X and Y column headers to describe your variables, for example “Hours studied” and “Exam score”.
- Enter one point per row. Paste two columns straight from a spreadsheet to fill hundreds of rows at once.
- Optionally fill in Group to color points by category, and Point label to name individual points.
- In Customize → Chart options, add a linear trendline for all points or one per group.
- Set axis titles and, if needed, fixed axis limits, then download.
Data format
| Column | Type | Notes |
|---|---|---|
| X | Number | Required |
| Y | Number | Required |
| Group (optional) | Text | Optional |
| Point label (optional) | Text | Optional |
Up to 5,000 rows. One row per point. Rename the X and Y headers to describe your variables. Fill in Group to color points by category; each group gets its own legend entry and, optionally, its own trendline.
Customization options
- Trendlines: ordinary least-squares regression for all points or per group, with equation and R² displayed.
- Marker size, opacity and shape per group. Transparency reveals overlapping points.
- Axis limits for X and Y, or include zero on both axes.
- Point labels to annotate notable observations.
- Gridlines on both axes to help read values.
Downloading and exporting
- PNG: best for slides, documents and the web. It supports transparent backgrounds, and you can choose standard, 2×, 4× or slide-sized (1920 × 1080) output.
- JPEG: a smaller file with a solid background.
- SVG: a true vector file. It stays sharp at any size and can be edited in Illustrator, Inkscape or Figma.
- PDF: vector PDF sized to the chart, A4 or US Letter. If your labels use characters outside the standard PDF fonts (for example Chinese, Arabic or emoji), the PDF uses a high-resolution image instead and tells you.
- CSV and .json: your data, or the whole chart with its settings, to re-open later.
When to use a scatter plot
- Exploring whether two measurements are related, such as height and weight or ad spend and sales.
- Spotting clusters, outliers and non-linear patterns.
- Comparing groups on two variables at once.
When to choose a different chart
- One variable is a category rather than a number. Use a bar graph or box plot.
- You need to show a third numeric variable. A bubble chart adds size as a dimension.
- The x-values are evenly spaced time steps. A line graph emphasises the sequence.
Common mistakes to avoid
Reading causation into correlation
A strong trendline only shows that the variables move together. Ice-cream sales and sunburn are correlated because both depend on sunshine.
Over-trusting R²
R² measures how well a straight line fits, not whether the relationship is real or important. A curved relationship can have a low R² even when it is very strong.
Hiding overplotting
With hundreds of points, many may sit on top of each other. Lower the marker opacity to see where data is dense.
Practical examples
Study time vs. scores
15 students’ hours studied and exam score with a trendline showing roughly 4.3 extra points per hour (R² ≈ 0.97).
Marketing spend
Monthly ad spend vs. revenue, grouped by channel, with one trendline per channel to compare their return.
Quality control
Machine temperature vs. defect rate. Label the outlier points with their batch ID.
How the trendline is calculated
The trendline is an ordinary least-squares (OLS) fit of y = a·x + b. It chooses the slope a and intercept b that minimise the sum of squared vertical distances between the points and the line. R² is the square of the Pearson correlation coefficient: 1 means every point lies exactly on the line, and 0 means x explains none of the variation in y.
The line is drawn only across the range of x-values you actually have. Extending it beyond your data (extrapolation) assumes the relationship continues, which may not be true.
Frequently asked questions
How do I add a line of best fit?
In Customize → Chart options, set “Trendline” to One line for all points (or one per group). The equation and R² appear in the top-right corner.
Can I plot several groups in different colors?
Yes. Type a group name in the Group column for each point. Each group gets its own color, marker shape and legend entry.
What’s the maximum number of points?
Up to 5,000 points per chart. For more, consider sampling your data or using a heatmap of binned values.
Can I use dates on the x-axis?
This tool expects numbers on both axes. For time series, use the line graph maker. For event dates, use the timeline maker.
Does the scatter plot support logarithmic axes?
Not currently. If your data spans several orders of magnitude, you can take the logarithm of the values in your spreadsheet before pasting them in, and note it in the axis title.
Your data stays private
Charts are drawn by your browser. Your data isn’t uploaded to our servers; it is kept only in this browser’s storage, for autosave and saved charts, until you clear it. See our privacy policy.