Dot Plot Maker
Here you can create a dot plot online: copy your data into the table above, select one metric variable and one categorical variable, and the dot plot is drawn for you. Every single observation is shown as one point, so nothing about your data is hidden behind a summary.
Depending on the field, the same chart is called a dot plot, a column scatter plot, a scatter dot plot, or an individual value plot. In numiqo it is listed as "Dot Plot (Individual Value Plot)" under Descriptive / Charts.
How to create a dot plot online
- Open the Statistics Calculator and paste your data into the table, one row per case.
- Choose the tab "Descriptive / Charts".
- Select one metric variable (the measured value) and one nominal variable (the group).
- Scroll down to the chart called "Dot Plot (Individual Value Plot)".
- Choose what to show on top of the points: mean, median, the group size n, and error bars.
- Download the figure as PNG or SVG, or copy it straight into your manuscript.
Use SVG when the chart goes into a paper. SVG is a vector format, so the figure stays sharp at any size and most journals prefer it over a screenshot.
What is a dot plot?
A dot plot shows one point per observation, arranged in a column for each group. Unlike a box plot or a bar chart, it does not replace your data with a summary: the reader sees how many measurements there are, where they cluster, whether the groups overlap, and whether a single extreme value is driving the difference.
A horizontal line marks the mean or the median, and optional error bars show the spread of the data or the precision of the mean.
Options of the dot plot
- Mean and Median — drawn as a short horizontal line through the column. Both can be shown at once; the median is dashed.
- N — adds the group size below each category label.
- Zero line — forces the value axis to include zero.
- Error bars — none, standard deviation, standard error, or 95% confidence interval.
- Orientation — columns running vertically or rows running horizontally.
- Size of the graphic — from small to extra large.
You can also add a reference line, and the chart title and the axis labels can be edited by clicking on them.
Which error bars should you use?
This is the choice reviewers notice, because the three options answer three different questions:
- Standard deviation (SD) describes how much the individual measurements scatter. Use it when the message is about the variability in your sample.
- Standard error (SE) describes how precisely the mean is estimated. It gets smaller as the sample gets larger, so it says nothing about the spread of the data.
- 95% confidence interval is the range that would contain the true mean in 95 out of 100 samples. Use it when the figure is meant to support a comparison between groups.
Whatever you pick, name it in the figure caption together with the group size, for example: "Points show individual measurements, the line the mean, error bars the standard deviation (n = 12 per group)." A caption that only says "error bars" is a common reason for a query from the reviewer.
How the points are arranged
Points that would fall on top of each other are moved sideways just far enough to stay visible, so no observation disappears behind another one and the width of a column shows where the values pile up. Points that do not collide stay in the middle of the column. The arrangement is calculated, not random, so the same data always produces the same figure, which matters when you have to reproduce a chart for a revision.
Dot plot, box plot or violin plot?
| Chart | Shows | Best for |
|---|---|---|
| Dot plot | Every single value, plus mean or median | Small samples, roughly up to 30 values per group |
| Box plot | Median, quartiles, whiskers, outliers | Medium to large samples, comparing many groups |
| Violin plot | The estimated shape of the distribution | Large samples where the shape itself is the message |
With few observations a box plot can be misleading: quartiles calculated from eight values look just as authoritative as quartiles from eight hundred. A dot plot shows the reader what the summary rests on. With several hundred points per group the opposite is true and a box plot or a violin plot is easier to read. A raincloud plot combines both ideas.
Why journals ask for individual data points
A bar chart with an error bar hides the sample size, the shape of the data and any outlier. Several journals therefore ask authors to show the individual measurements for continuous data with small samples, and the dot plot is the standard answer to that request. It is the chart type that scientific papers describe as a column scatter plot or scatter dot plot.
Frequently asked questions
Is a dot plot the same as a scatter plot?
No. A scatter plot puts two metric variables against each other to show a relationship. A dot plot puts one metric variable against groups to compare them.
What is a column scatter plot?
It is the same chart under a different name. "Column scatter" is the term used in GraphPad Prism, "individual value plot" the term used in Minitab, and both describe the dot plot on this page.
Can I show mean and median at the same time?
Yes. Tick both under "Show". A visible gap between the two lines is a quick sign that the data are skewed.
How many groups can I compare?
As many as your grouping variable has categories. With a second nominal variable the columns are nested, so you can show two factors in one figure.
Can I use an ordinal variable for the groups?
The dot plot is drawn for a metric variable together with a nominal variable. If your grouping variable is stored as ordinal, change its scale level to nominal in the data table.