Oct 14, 2006
A warm welcome to readers of Science. (Junk Charts is selected as "Best of the Web" this week. Also thanks to Mitchell for the nice write-up.)
Racetrack graphs was a novelty item here some time ago. They made an appearance in the October issue of Wired Magazine, known for its design. We have already discussed information distortion in such charts.
This chart fails the self-sufficiency test, forcing readers to read and interpret the data labels, and to ignore the racetrack construct.
Graphical elements applied as cosmetics? Charts sacrificing data integrity for entertainment? This takes us back to our previous discussion: can good charts be entertaining? Now flipped over: can entertaining charts be good?
Reference: "Good, Green Livin'", Wired Magazine, 10/2006.
Another problem with this graph is the the choice of units determines the graphic point.
Cloth diapers look bad because the green line is so much longer than the other lines. But the green line is gallons of water -- so the units are, obviously GALLONS. These are compared with MEGAJOULES and POUNDS. Can these units be meaningfully compared? Not by me, at least.
Posted by: zbicyclist | Oct 14, 2006 at 02:32 PM
At least the original racetrack graph you mention was based on values that are percentages of a whole. This graph seems to make the length of the green line completely arbitrary.
Posted by: cscmike | Oct 14, 2006 at 04:34 PM
zbicyclist: and it's so unnecessary. There are three racetracks in three categories, so not a thing would have been lost in presenting one racetrack each for water, solid waste, and energy, with the compared values being disposables, cloth homewashed, and cloth service washed.
To summarize: it didn't have to be a graph, a table would have done as well; it didn't have to be a racetrack, a bar graph would have been better; and it didn't have to be those racetracks, a switch of categories would have been better.
What an execrable choice of graphic.
Posted by: derek | Oct 16, 2006 at 07:30 AM
I completely agree with derek, they obviously only did this to make something that would look good. But charts can be good and entertaining, see e.g. these ones from Princeton. While there are some problems (e.g., scaling images in both directions at once, thus creating an effect that grows with the square of the number), they are informative and fun. Are they biased? Sure, but that's their whole point.
Posted by: Robert Kosara | Oct 17, 2006 at 03:52 PM