Jun 26, 2007
Xan G. tells us that these "inconsistent pie charts ... make [his] head hurt". The dizzy array of colors is unfortunate, especially when "Application" gets a medium blue in three of four pies but an orange-red in one of them. Just like the baby names charts, it's important to keep the background constant when constructing small multiples.
We cite from the horse's mouth:
The goal of this section was to uncover any [software development] task that might be overlooked [by these startup companies]. When writing a software product, the tendency is to focus 100% on the application. Items like support, marketing, and especially billing never cross your mind.
The junkart version below is designed to bring out this one message: that Blinksale has distinguished itself from the rest by having spent more time developing code for purposes other than the application itself.
I removed the raw counts of lines of code and focused only on the relative proportions. The former does nothing to argue the author's case.
The pie charts fail our self-sufficiency test. The reader must rely on the data table and data labels to understand the chart. If removed, the key message is obscured.
Source: "Web App Autopsy", ParticleTree, June 2007.
I am not sure if you should use a line chart to display nominal data. Also, it is not easy to identify each series.
And there is (again) the "fun factor". Should be possible to have a "fun" chart without compromising the message.
Posted by: Jorge Camoes | Jun 27, 2007 at 03:45 AM
As Jorge points out, a line chart is inappropriate, since it implies some kind of sequential relationship between adjacent categories. A bar or column would be better. My first impulse is to make a clustered bar, probably with startup as category and purpose as series, but seeing the initial attempt may reverse startup and purpose.
In fact, a simple table would be sufficient for this data, although it gets messy if you want to display percentages and actual numbers.
And I'll point out that it's very difficult to mix "fun" with information display. "Fun" seems to include 3D effects. clip art, outrageous colors, and dazzling effects.
Posted by: Jon Peltier | Jun 27, 2007 at 08:57 AM
I'd be partial to a stacked bar chart with application as the bottom bar (so its bottom edge remains consistent for comparability -- and getting the main message across).
Posted by: John Johnson | Jun 27, 2007 at 10:05 AM
I agree with Jorge & Jon about the inconvience of the line chart. I had a go at a mekko chart wich partly helps in taking into consideration the difference of scale in terms of total spent between the 4 companies. Lacking a good Mekko tool to create a good-looking graphic though.
Posted by: Bernard Lebelle | Jun 27, 2007 at 10:59 AM
Line chart? for nominal data? The horror!!!
Surely, this was just done to see if any of us were awake!
Posted by: zbicyclist | Jun 27, 2007 at 09:37 PM
The revulsion to the line chart is unexpected. This construct is used frequently by statisticians, usually called a "profile plot". I'll put my thoughts in a follow-up post.
Posted by: Kaiser | Jun 27, 2007 at 11:40 PM
I used to go along with the idea that lines have to be about time series or some other ordinal scale, but seeing the utility of "parallel coordinates" displays cured me.
Posted by: derek | Jun 28, 2007 at 05:36 AM
Bernard: you might have more luck finding software to produce "mekko" charts if you look for their proper name (rather than a named coined by a company trying to sell them) - mosaic plots. R has a number of packages that produce mosiac plots, and there are a few standalone programmes too - mondrian is written in java and fairly easy to use.
Posted by: Hadley Wickham | Jun 29, 2007 at 02:37 AM
I was thinking of a 2D mosiac plot, too, when I first saw these pie charts. Here's one for the charts featured here:
Posted by: Xan Gregg | Jun 29, 2007 at 08:26 AM