KavaChart in Action - Examples and Documentation - AlaCarte

[ Back ] Specialty Applet Collection - Radar Charts

Kiviat, Polar, Radar, or Spider charts give you a visual comparison of several items, usually without regard to a specific numeric scale. For example, lets say you wanted to compare several individuals to see who was best suited for a particular position. A spider chart lets you rate several attributes against an ideal, and get an instant visual comparison of any individual against the ideal candidate.

KavaChart AlaCarte's Specialty applet collection includes the applet "com.ve.kavachart.applet.polarApp" for drawing Radar charts.

Use the source code to this page as a guide in creating your own spider charts. You can also use the KavaChart ChartWizard to design your own spider charts.
 

A simple spider plot has one "spoke" on the axis for each item compared. The user-defined axis labels are positioned at the end of each spoke, and numeric labels are drawn on the vertical spoke at the 12 o'clock position.

With a Kiviat chart, all the values are typically normalized to a specific range, generally from 0 to 1. The shape of the various datasets can be compared with the ideal at a glance, which is the goal of this kind of chart. Each dataset should have the same number of items, and each item's value should fall within the same range.

KavaChart spider plots are implemented with the applet "polarApp", which can also be used for polar charts.

 

This spider plot compares some elite NBA teams. At a glance, we can see which teams take the most shots per game, which team has the best free throw percentage, and so on.

Since these values all have different ranges, we normalized the values to a number between 0 and 1, and then set the axis to scale from -0.5 to 1 to keep the dataset shapes roughly circular.

Kiviat charts can also display markers at each points, and can toggle the visibility of the plot lines for a sort of scatter plot display. This is more effective when the axes are reduced with the "numSpokes" param, or when the axes are hidden by setting the color to "transparent".

You can also fill the areas in the spider plot, but it's a good idea to use partially transparent colors if you do. This will let you see lower layers through the upper layers.

 

With a much larger number of points and a fixed number of axis spokes, a spider plot takes the appearance of a radar plot. KavaChart's polarApp always distributes the points in a circular distribution around the center of the plot, but you can vary this behavior in a variety of ways with a little bit of Java programming. See the "contrib" item "dailyPolarApp" for an example applet that uses this object to create a 24 hour "clock" of values.