Radar Chart Maker
Make a radar chart with several series overlaid on the same axes.
What this tool does
A radar chart plots several variables on axes radiating from a centre, joining the values into a shape. It is useful for comparing a few items across the same set of measures - and poor at almost everything else, which is worth knowing before you choose one.
Draws multiple series on shared axes with a scale maximum you set explicitly rather than one inferred from the data.
Warns when a series has the wrong number of values, because a missing value plots as zero and looks exactly like a genuine low score.
How to use it
- List your axes, one per line.
- Add each series as name, then one value per axis.
- Set the scale maximum so it matches your rating scale, or press Fit to data when the values are not scores out of a known maximum.
- Compare the series totals beneath the chart - the shape is read as area, which the eye judges badly, so the totals say plainly which series is bigger.
- Export as SVG or PNG.
Why use this chart maker
Radar charts compare profiles across several measures, such as skills or product ratings. Setting the scale yourself avoids misleading charts.
Why we built it
We built the radar chart maker to overlay several series on the same axes, which is the one thing a radar chart does better than a bar chart. Compare exact values with the grouped bar chart maker and show shares with the pie chart maker.
For a neighbouring job there is also the Doughnut Chart Maker and the Box Plot Maker.
Who this tool is for
- Managers and HR comparing skills or scores
- Product reviewers rating features
- Students comparing profiles
When to use it
- Comparing skills, products or teams.
- Player or character stat sheets.
- Survey scores across categories.
- Presentations with multi-axis profiles.
What radar charts are genuinely good at
One thing: comparing two or three items across the same handful of measures, where the shape of the difference matters more than the exact values.
That is a real use. A product compared against a competitor across six attributes shows immediately where each is stronger, and the eye picks up the shape difference faster than it reads a table.
Outside that, a radar chart is usually the wrong choice, and the reasons are worth knowing because they are structural rather than cosmetic.
The three problems
Area misleads
The enclosed area grows with the square of the values, so a series scoring twice as much on every axis encloses four times the area. The eye reads area, so differences look far larger than they are.
Axis order changes the shape
Reorder the axes and the same data produces a completely different outline. Nothing in the data determines the order, which means the shape is partly an artefact of an arbitrary choice.
Comparing across axes is meaningless
Each axis usually measures a different thing on a different scale. A value of 8 on Speed and 8 on Cost sit at the same distance from the centre while having nothing in common, which invites conclusions the data does not support.
If the exact values matter more than the overall shape, a grouped bar chart will serve you better. Radar charts are for impressions, not for reading numbers off.
Making one that works
- Three to six axes. Below three there is no shape; above eight it becomes unreadable.
- Use the same scale on every axis and set the maximum explicitly, so the chart cannot silently rescale when the data changes.
- Two or three series at most. Overlaid polygons obscure each other quickly.
- Keep the axis order fixed across every chart in a set, or the comparison between charts is meaningless.
- Normalise your measures to a common scale before plotting if they are not already.
What you give it, and what you get back
| Item | Details |
|---|---|
| Input | Axis names and values for one or more series. |
| Scale | A maximum you set explicitly. |
| Output | Overlaid series on shared axes. |
Issues you may run into
| What you see | Why it happens | What to do |
|---|---|---|
| A series looks wrong | It has the wrong number of values. | The warning identifies it. |
| Differences look exaggerated | Area grows with the square of values. | Compare values, not areas. |
| A row is skipped | It does not match the expected format. | Check the message listing unreadable lines and fix them. |
Assumptions and limitations
- Enclosed area grows with the square of the values, which exaggerates differences.
- The shape depends on axis order, which the data does not determine.
- Values on different axes are not comparable even though they are drawn at the same radius.
- Practical maximum is around eight axes and three series.
- Negative values cannot be plotted - the scale starts at zero.
Frequently asked questions
What is a radar chart used for?
Comparing two or three items across the same set of measures, where the shape of the difference matters more than exact values. A product against a competitor across six attributes is the classic case.
When is a radar chart a bad choice?
Whenever the exact numbers matter, whenever you have more than about eight axes, or whenever the axes measure things on genuinely different scales. In all three cases a grouped bar chart communicates better.
How many axes should a radar chart have?
Three to six. Below three there is no shape to read. Above eight the polygon becomes too complex to interpret and the labels start colliding.
What is the difference between a radar chart and a spider chart?
None - they are the same chart under different names. Web chart, star chart and polar chart are also used, though a polar chart sometimes means something different in scientific contexts.
Why does the shape change if I reorder the axes?
Because the outline is drawn by joining adjacent axes, so adjacency determines the shape. Nothing in the data specifies an order, which is one of the genuine weaknesses of the format - keep the order fixed across a set of charts.
Can I compare more than one thing?
Yes, and that is the main use. Two or three series work well overlaid. Beyond that the polygons obscure each other and the chart stops being readable.