On debatable infographics

Below is an infographic of the length of time each candidate spoke at the Democratic debate earlier this month.

No one speaks as quickly as Clinton, I guess. Source: Politico

Regardless of your political preferences, we can hopefully agree that this visualization misses its mark when it comes to intuitive comprehension. To preface, I found this via a Reddit user who was frustrated looking at this.

Infographics have become a trend any time there is some data to display, as they are theoretically easily consumable and attractive. While there are certainly ones that do that, many unfortunately become examples of how not to create an infographic, often because flashiness is prioritized over legibility.

Politico’s infographic above represents limited data — speaking time per candidate. We’re comparing three numerical values, each associated with a different person. Instead of juxtaposing these values in closer proximity so that the reader could easily determine how they relate to each other, this visualization has three separate odometer-like graphics that are individually taking up a lot of space. There was really no reason to go with something like an odometer, as this infographic has nothing to do with speed. But looking at this, it’s at least clear that this is ordered from most speaking time to least, though Martin O’Malley’s graphic being on a new line doesn’t help to identify this.

For whatever reason, O’Malley’s is apparently on a different scale ending at 40, which is misleading… though it could just be a typo. And therein lies another problem: the difficulty of estimating values along arcs/circles. Upon closer inspection, it does appear that O’Malley’s arc is out of 40, and you can also determine this by reading the line pointing slightly over halfway, though this requires more second guessing and double takes than it really should and the differences in times between the candidates are not as obvious or transparent as they should be.

The information hierarchy is also basically non-existent, as the graphics, without context, just have small labels specifying the important information. There are also, for yet another unknown reason, three different ways of writing out the times, which certainly doesn’t boost my confidence in what this infographic or article is trying to relay.

With the semi-circles, I wonder if they’d originally meant to go in the direction of a clock, which would better relate to the subject of the data but hinges precariously on including graphics for the sake of pretty pictures, unless done tastefully (which seems dubious at this point).

Honestly, Politico could have just gone with a simple, clearly labeled bar chart, perhaps stacked to show the candidates’ speaking times as fractions of each other. (Which, in looking at the article, they do have for the percentage of time the candidates spoke, compared across the debates so far. Why they went an entirely different design route for the speaking times is beyond me.) In short, there is little data here to display, yet they have somehow managed to do it in a quite obtuse way.

There’s also the difficulty of estimating values along an arc like this. Upon closer inspection, it does appear that O’Malley’s is out of 40 for some unknown reason, though