What it is
A bubble chart extends the scatter plot: x and y place each item, and the bubble’s area shows a third value such as population, revenue or count.
It is best for a modest number of named items — countries, products, planets — where the eye can take in position and size together.
✓ Use it when
- You have three numeric measures for each of a few dozen items at most.
- The third measure is a size-like quantity: population, sales, volume.
✗ Avoid it when
- Precise comparison of the third measure matters — area is judged less accurately than position.
- There are hundreds of items — the bubbles hide each other.
Tips for a better chart
- Scale bubbles by area, not by radius. Chart Buffet does this for you.
- Make bubbles translucent with a thin outline so overlapping ones stay visible.
- Label the bubbles that matter; let the tooltip carry the rest.
The maths
An honest bubble has area proportional to the value: A ∝ v, so the radius grows as the square root, r ∝ √v. Scaling the radius directly would make a value twice as large look four times as big.
People still tend to underestimate the size of large bubbles. Perceived area grows more slowly than real area, following Stevens’ power law with an exponent below one for area (Stevens 1957), so bubble sizes are best read as rough comparisons.
A little history
Proportional circles appear in nineteenth-century thematic maps, and statisticians scaled scatterplot points by a third variable well before computers (Friendly & Denis 2005). The bubble chart’s most famous moment was Hans Rosling’s 2006 TED talk, in which animated bubbles of the world’s countries raced across a chart of income against life expectancy (Rosling 2006).
Questions
Which column sets the bubble size?
Choose it under “Bubble size” in the Data section. Chart Buffet picks the third numeric column by default.
References
- Stevens, S. S. (1957). On the psychophysical law. Psychological Review 64(3), 153–181.
- Rosling, H. (2006). The best stats you’ve ever seen. TED talk.
- Friendly, M. & Denis, D. (2005). The early origins and development of the scatterplot. Journal of the History of the Behavioral Sciences 41(2), 103–130.