What is a radar chart?

It is also known as a spider chart, web chart, star chart, cobweb chart, Kiviat diagram, polar chart, or irregular polygon. The inventor of this family of charts was a German man named Georg von Mayr, who published the first known radar chart in 1877.

As a simple example of a use for radar charts, imagine your favorite brownie. There are many factors, or variations, that make up a brownie: chewiness, chocolatey-ness, the addition of nuts, and other ingredients such as cranberries, as well as the crust, moistness, and density.

A radar chart for brownies would have a ‘spoke’ for each factor, and the mark on the length of the spoke would indicate the measurement for that variable. For instance, your mom’s brownie might rank highly on chewiness, but she doesn’t include nuts or other extra ingredients. Your favorite bakery might rank differently, with the inclusion of walnuts and a cakier texture. Then, a line is drawn from each ranking for each variable, so it looks like a spider web.

When should radar charts be used?

Radar charts are most beneficial when there are a few items to compare. This can be done by overlaying different product information on the same chart or by having several charts displaying the same radii but analyzing different products. For instance, you could compare your mom's brownies, the bakery's, and your neighbor's by having charts next to each other, or instead, the measurements all mapped on the same chart.

In terms of business use, there are many possible uses. Consider skill analysis for staff members. They could be assessed in terms of communication, problem-solving, teamwork, ability to meet deadlines, punctuality, and technical understanding. A radar chart immediately shows where staff are assessed in comparison to their colleagues.

Another use for radar charts in businesses is to manage quality improvement, as a radar chart can be used to display performance metrics.

• Use a radar chart when:
• There are multivariate observations
• There is an arbitrary number of variables
• You need to identify outliers
• You need to make comparisons across products or services
• Data sets are small or moderately sized

When creating a radar chart, there are a few best practices:

• Variables should be arranged in some meaningful order
• More than three series should be presented on their own radar charts
• Don’t use too many variables or the chart risks becoming confusing
• If there are multiple data series, the filled-in color should be transparent

Advantages of a radar chart

Outliers and similarities are easy to see

The most significant advantage of using a radar chart is that outliers are immediately visible. Any metric or variable that is very different from the others on the chart or in a set of charts is very obvious. Commonalities are also easy to assess, particularly if they are plotted on the same chart.

Disadvantages of a radar chart

However, there are also multiple disadvantages and difficulties involved in using a radar chart that will be discussed in further detail below.

Hard to judge Radii length

The distances on the radii are difficult to judge visually. While people will know and understand that one variable is longer or shorter than another, this can be difficult to quantify. In this instance, using concentric circles on the chart can help to make the radii length easier to judge. Consider also using a line graph in this case, if understanding the length is important.

Radar charts can distort data

Once all the measurements are lodged on the chart, the area is filled in. This can distort data, as the area shaded in becomes a visual judgment of size and positive performance. Also, if a chart has five variables measured up to 100 on the radii, a chart that results in a measurement of 90 is 10 percent bigger than a chart that adds up to 82.

Radar charts can create connections where there is none

If you have five variables on a chart, and they are all labeled on the radii, it could be tempting to think there is a relationship between the juxtaposed measurements. However, using the brownie example, there could be no relationship between the texture and level of chocolate flavor, as they are two separate variables.

Radar charts can cause occlusion and confusion

If there are too many variables or too many data series, the chart can become confusing. This problem is compounded if multiple data series are plotted on one chart, and data points can become occluded. While keeping the color shading transparent can help stop occlusion, if multiple series hit the same variable measurement, it becomes more difficult to extract a clear picture of data.

Similarly, psychology comes into play in regards to shapes. Humans recognize and can discern data in shapes like squares, circles, and triangles. The random nature of radar chart shapes makes them less useful than known and quantifiable shapes.

Alternatives to radar charts

Below are some common alternatives to radar charts.

Parallel coordinates plot

The parallel coordinates plot allows for the comparison of several features using a set of numeric variables. Each variable is plotted onto a vertical line, and then the multiple variables across the chart are joined together. This is ideal for use if there are many items all measured on the same scale that need comparison. For instance, if you were comparing 30 staff members rather than two, a parallel coordinates chart would be an effective way of showing that.

Glyph plots

These plots are similar to radar charts in that the variables are represented by spikes that are proportional to the data. While glyph plots map data attributes onto a graph, much like radar charts, they are used in very different scenarios. A Glyph plot can have more—or fewer—variables than a radar chart and is often used in multiples to show patterns over time.

Line graphs

The simple alternative to a parallel coordinate plot, a line graph can show the rankings on a range of variables. While not as visually pleasing as a radar chart, it can be an effective way to show differences between a small number of data sets.

Harvey balls

Harvey balls are circular characters that communicate qualitative information. They look like circles with sections cut out of them. For the brownie example, there would be six different balls, each representing a separate variable. These may be better to use when there are fewer variables—no more than four or five, in order to keep a simple and easily understood chart.

Bar charts

Bar charts, or a stacked bar chart, could be another option for displaying data. A stacked bar chart, in particular, could be effective as it will clearly show the overall difference in totals of variables. This also could be suitable for a number of data sets.

Bar charts are effective for displaying data with small multiples, an approach developed by Edward Tufte. Individual series of data are plotted onto mini bar charts and displayed alongside other data series, which are easy to interpret.

Also under bar charts, lollipop charts could be considered an option. A series of thin bars with dots on the end are another easy way to present information that could otherwise be on a radar chart.

The future of radar charts

While radar charts look interesting, they are ultimately not the best solution for presenting data. There are problems with interpreting the information, which can easily become confusing for many. Many other charts do a far more effective job in presenting multivariate data.