Unless you’re a data scientist or have another job that commonly uses data visualization, selecting the right way to represent information can feel overwhelming. Software like Excel can stuff any type of data into almost any type of chart—but which one is the right one for the story you’re trying to tell?
Bar charts and pie charts are two common ways of representing data, and while they can sometimes be used interchangeably, they don’t make sense for every type of dataset. Here’s how to choose between a bar chart and a pie chart.
When to use a bar chart
Bar charts have two axes: categories (months of the year) on one and numeric value (the number of visitors to a national park) on the other. So you’re comparing the total visitors in January to those counted in other months, and perhaps observing which months have the most visitor traffic. Bar charts can be oriented horizontally or vertically.
Bar charts can also be used to combine datasets or show how subcategories compare—for example, the number or percentage of votes each candidate for office received broken down by voter age group.
When to use a pie chart
Pie charts, on the other hand, show categories as a proportion of a whole. You can think of the circle as the total—100%—and the width of each slice is determined by how much of the total it represents relative to the other slices. Using a similar example, a pie chart could show the percentage of votes each candidate received out of the total number of votes cast by all voters.
Pie charts are really only useful when the individual data points you are including add up to a whole, and that whole in itself is part of the story. If you’re trying to convey how a part is related to a whole—for example, what percentage of your budget is spent on rent—a bar chart may not do that as clearly.
When to choose a bar chart over a pie chart
There are a few scenarios in which, despite the data representing parts of a meaningful whole, a pie chart would be a poor choice:
- There are too many categories. If the pie slices are too narrow, it’s difficult to determine their relative size. Plus, the chart becomes crowded and labels and annotations are hard to read.
- Slice sizes are too similar. If pie pieces have proportions that are close but not equal, the relative size may be difficult to interpret. At a quick glance, and without clear labels, pie charts may not clearly convey the exact proportion a slice represents.
- You have multiple datasets. Pie charts can only represent one dataset that, again, adds up to a whole. They cannot compare salaries for certain job categories further divided by gender, for example.
While pie charts are especially good at showing percentages relative to a whole, bar charts can also be useful for showing percentages rather than absolute numbers, especially when the above limitations exist or if categories aren’t mutually exclusive—for example, a survey that allows users to select multiple answers or categories for a single question.
In general, use a bar chart when you’re comparing category to category, and a pie chart when you’re comparing part to whole. But a bar chart may still be better for part-to-whole visualization if you have a lot of parts (more than five) or if the proportions of the pie slices aren’t easily identifiable (such as multiples of 1/3 or 1/4). It doesn’t go the other way: pie charts aren’t useful for category-to-category comparison if the whole doesn’t have any meaning. The bottom line: bar charts are just more versatile than pie charts.
Read the full article here