Commonly used on maps, and x/y-axis plots, or no plot at all, bubble charts communicate the raw count, frequency, or proportion of some variable where the size of the bubble reflects the quantity. Color-coding bubbles can represent a further categorization of the variable being graphed. It is also possible to add another dimension by showing the movement of bubbles over time (referred to as a motion chart).
The difficulty with bubble graphs is in their interpretation. While they can give a quick comparison of values of your data, they are not as well suited for accurate or precise determination. Some software graphs bubbles by volume, not diameter, further complicating comparisons. Other variables become difficult to graph by bubbles if the range is too large and some too small. In such cases, bubble charts can become a challenge to interpret.
This blog post on the Scientific American website illustrates how area can be difficultly and sometimes incorrectly interpreted.
There are many more examples of poor usage of the bubble chart so be careful in looking at examples. The two below are good representations of how the bubble chart can be used.
Frequency of words at national conventions
This chart shows the frequency of words used at the National Conventions. The colour within circles reflects the political party.
Relationship between health and wealth
This example shows the relationship between health (represented by expected years of life at birth) and wealth (represented by GDP per person) in each country in the world. The size of the bubbles in the graph represents the population of each country.
Advice for CHOOSING this option (tips and traps)
This option is particularly useful for conveying a large amount of numeric information quickly. Ensure that your audience is visually literate.
Make sure you have at least three sets of linked variables that you want to investigate.
Ensure the data collectors are all collecting data in prescribed formats and in software that allows for upload to common bubble chart creation sites.
Make sure it is clear what each element of the chart means – color, circumference, how it fits on the scale, etc. otherwise meaning can be lost.
Advice for USING this option (tips and traps)
Know the capabilities of the software available. Features can vary between software packages.
Collect data in prescribed formats and in software that allows for upload to common bubble chart creation sites.
This tool is useful for showing situations but not for answering questions on ‘why?’ and ‘how?’. This option will need to be supplemented with others if the evaluation is investigating the reasons behind problems or potential solutions.
Position the center of the bubble over the center of the geographic area it is associated with, if applicable.
This option requires some understanding of statistics to draw bubble charts manually. For example, the area of a circle in bubble charts is proportional to the square of the radius. To get a properly weighted scale, take the square root of the third metric rather than the metric itself. Alternatively there are software programs, including Excel 2013, that will draw them for you automatically.
How to Make Bubble Charts: This tutorial from Flowing Data, provides a detailed guide on creating bubble charts using the software provided.
Gapminder: This website allows you to enter data or use the data sets that are available on site on create bubble charts.
Other ways to compare sets of values
Illustrating the main features of the distribution of a data set in a clear way.
Presenting a frequency distribution of quantitative data in a graphical way.
Using a target line to show progress to date, often with levels of performance graphed in the background.
Deviation bar graph
Aligning two bar graphs along their spine to compare the shape of their data sets.
Plotting two or more dots on a single line for each category being compared.
Positioning several small graphs with the same scale in a row for easy comparison.