9.2.5: Radial Plots (Circular Charts like Radar or Polar Charts)

Radial Plots (Circular Charts like Radar or Polar Charts)

Radial plots refer to a family of charts that use a circular layout to represent data. Common examples include radar charts (also known as spider or web charts), polar area charts, and sunburst or radial tree charts. Here we focus on radar charts, as they are a frequently used radial visualization for multivariate data [12]. A radar chart has multiple quantitative axes arranged radially like spokes on a wheel; each axis represents a different variable, and a data point is plotted by connecting the values across all axes, forming a polygon shape. For example, to compare the performance of three departments across five metrics (say, revenue, growth, customer satisfaction, market share, and profit margin), one could use a radar chart with five axes. Each department’s metrics form a polygon – by overlaying them, you can see which department is excelling in which areas and their overall profiles.

When to use a radar chart: Radar charts are useful for comparing multiple entities on several variables at once, especially to identify relative strengths and weaknesses or patterns in profiles. They are often seen in performance evaluations, for instance: an employee’s skills across different categories (communication, technical skill, leadership, etc.) might be plotted to show a skill profile. Similarly, product comparisons (e.g. comparing smartphones across battery life, camera quality, screen resolution, etc.) can be illustrated with radar charts, letting you see at a glance which product is better in which dimensions. Radar charts also allow grouping – for instance, you might plot one polygon for this year’s values and another for last year’s values, to see changes in a multi-dimensional performance space.

The advantage of a radar chart is its holistic view: you can examine overall shapes. If one polygon’s shape vastly extends in certain directions compared to another, it shows divergence on those specific metrics. If it’s uniformly larger, that entity is better on all fronts (rare but easy to spot). If the shapes intersect, each might have different spikes.

However, radar charts come with significant caveats. They can be misleading or hard to read if not used carefully. First, the eye has trouble comparing lengths at different angles – comparing values on non-adjacent axes is not as straightforward as comparing bars on a common baseline. Second, if you have many variables (axes) or many series (polygons), the chart becomes cluttered very quickly. Experts generally advise using radar charts for no more than ~5–8 variables and a handful of series. If you exceed those, a radar chart can turn into a jumble of overlapping shapes and labels. Third, the variables in a radar chart are implicitly treated as if they form a cycle (since the last axis connects back to the first around the circle). This means the sequence of variables can affect the perception of the shape. If your variables have no natural order, the arrangement around the circle might be arbitrary and potentially misleading (viewers might perceive adjacent axes as related just because they’re next to each other on the chart). To mitigate this, one should order axes logically (e.g. group related metrics next to each other).

Radar charts should also be used with caution because they emphasize pattern recognition over precise values. They make it easy to see which metrics are “high” or “low” for a given entity relative to others, but reading exact numbers is not their strong suit. If precise comparison of one metric across many categories is needed, small multiples of bar charts or a table might be clearer. In fact, some experts critique radar charts heavily and prefer alternatives like parallel coordinate plots for multivariate data or just separate bar charts. Yet, when used in the right context – few variables, focusing on relative patterns – radar charts can be a visually engaging way to show a multivariate profile. They often appeal in executive or marketing reports due to their “snapshot” feel (e.g. a brand perception radar chart across dimensions like “innovative,” “trustworthy,” “value for money” can quickly show a brand’s strengths). Tools like Excel, Tableau, and Python’s Plotly can all create radar charts (Excel calls them spider charts). One just must be mindful of their limitations.

In summary, radial plots like radar charts are best for comparing entities across multiple attributes in a qualitative way. Use them to highlight which dimensions an entity is above or below others, or to show an overall pattern (e.g. a balanced shape vs a lopsided one). Avoid using them for large numbers of variables or where exact data reading is required. With these caveats, radial charts can effectively complement your visualization toolkit for specialized cases.