Showing the spread rather than the average.

The chart that shows whether the average you have been quoting describes anyone at all.

Drawn as a histogram, density plot, box plot, depending on the data.

Satisfaction is bimodal, not average

A mean of 3.94 describes almost nobody in this distribution

Rendered from the project Customer survey — Q2, through the same compiler the product uses. Not a picture of one.

What it is for

A distribution chart answers what the values look like as a population rather than as a summary. It is the form that catches bimodality, skew and outliers, all of which are invisible in a mean and all of which change what the mean means.

It earns its place most often when a single number has been driving a decision. Order values, payment delays, appointment lead times and satisfaction scores are all routinely bimodal, and a bimodal distribution has an average that sits in the valley between the two humps where nobody is.

Reach for it when

You are about to quote an average and have not looked at the spread
There is reason to suspect two populations mixed together
Outliers might be carrying a total, and you need to see how many and how far out
The reader's question is what is typical rather than what is the sum

Use something else when

There are too few observations for a shape to be meaningful, usually under about thirty
The audience needs to compare specific categories, where bars answer directly
The values are categorical rather than continuous

How this form misleads

Bin width chosen to make the shape you expected

Wide bins smooth a bimodal distribution into one hump; narrow bins turn ordinary noise into structure. The same data supports both pictures, and neither is a lie in itself.

Choose the width from the data rather than from the story, state it on the chart, and check that the shape survives a wider and a narrower version before drawing a conclusion from it.

Clipping the outliers off the axis

The long tail is frequently the finding. Cutting the axis at the 95th percentile removes exactly the observations that are moving the total.

Keep the full range, or break the axis visibly and say how many observations are beyond the break and what they add up to.

A box plot for a general audience

It is a dense, precise summary that requires the reader to know what the box, the line and the whiskers each mean, and it hides bimodality entirely — two distributions with completely different shapes can produce identical boxes.

Use a histogram where the audience is not statistical, and keep the box plot for comparing many groups at once where the shape is already established.

What this will not draw, whatever you ask for

The agent picks an intent from a closed list and a single compiler renders it. These forms are rejected at the schema layer rather than discouraged in a style guide, which is the difference between a rule and a preference.

3D anything: Perspective makes two equal values look different. There is no case where the third dimension carries data.
Gauges and speedometers: An enormous amount of ink for one number, and the dial implies a range that is usually invented.
Radar charts: Area scales with the square of the values and the shape changes entirely if you reorder the axes.
Word clouds: Size encodes frequency, which is the least interesting thing about a word, and nothing is comparable.
Pie charts above five slices: Angles are hard to compare, and beyond five slices the labels leave the chart anyway.
Per-series arbitrary colour: Colour means something specific in this product. A series that picks its own hue breaks the meaning everywhere else.

Make one from your own data

You do not choose the form. The analysis works out which question your data can answer and picks the chart that answers it, then tells you why.

No account needed to start. You only pay when you like what you see.