Originally published at: http://boingboing.net/2016/10/17/wtf-visualizations-names-and-s.html
…
3 Likes
Those hurt my brain to look at.
14 Likes
The other two are so much worst than the first one 
5 Likes
Gag.
One name for these bozos - .Edward Tufte.
5 Likes
I’m lost. Someone needs to make a visualization of the good to bad ratio for data vizualizations, separated out in to “good, bad, other, and yes”.
18 Likes
I’m disappointed to not see the “What Donald Trump has Lied About Today” visualizations on here.
24 Likes
After seeing how shitty the first two were I’ve avoided them. Factoids randomly connected by a cross between drug-addled spider and a clueless Freshman comp sci major is not a visualization.
10 Likes
They aren’t “visualizations”. The visuals contain NO information and are completely irrelevant. You can mix up text and it makes as much sense as anything else. It is more or less pointless. You may as well just make a list. I think it is done this way to appear “smarter”.
As for these graphs - eh the first one doesn’t really make it easy to see the percentages. The 2nd one - uh - need more info. Down from what? Last year? The peak in the 90s?
2 Likes
And when you have people who understand the data, but have limited knowledge of graphic design, you get things like this:
17 Likes
Complete with slurs! So sad.
9 Likes
To be honest, I didn’t mind the second one so much. It all depends on the context: what was the visualization trying to say?
As a comparison of the incidence of preterm births among different populations, it’s not actually that bad. It makes it clear that non-hispanic black women have a higher rates of preterm births. From that we can start asking why this is so.
It seems wrong because a pie chart should add up to 100%. But it does, in a more subtle way: Normalized to the different population sizes, what percentage of preterm births belong to each group? The percentages in the text the give additional info of the percentage preterm births within each group.
The “correct” way would have to be a bar chart, which would be fine, but not actually provide any better information.
The “seems correct but would actually be terrible” way would be to not normalize the groups, and show a pie chart the actual numbers of preterm births, categorized by group. This would be bad because if, say, the population were 90% white, the fact that non-hispanic blacks have higher rates of preterm births would be lost.
2 Likes
I’m having flashbacks to Molecular Biology now. All the illustrations in all the books were ridiculous. I realize there’s no way to illustrate this stuff, but still, the illustrations were ugly and reductive to the point of uselessness.
2 Likes
Lots of the graphics seem to have been picked because their data don’t match their representations.
For example, in the nested-circles graphic from the topic, there are two circles, of different sizes, both labeled “down 64%.”
As for graphics where the visuals contain no information and are completely irrelevant,
This one is more of a venial design sin, where the Trump lies “charts” are deadly.
4 Likes
Most of the bad visualizations I see simply lack labeling. The trends and proportions can be seen but without any idea what they mean.
2 Likes
Right, here the graphic is basically just space-consuming decoration for what would otherwise be a short bullet list. It isn’t ‘data visualization’ in any real sense.
It would be cool if they provided some actual commentary for the graphically clueless.
2 Likes
No it doesn’t. It’s just flat-out wrong. It was designed by somebody who didn’t understand when they went over pie charts in third grade. I challenge you to figure out what those numbers even mean.
6 Likes
Plenty of the ones in the inexplicably-lauded ‘Information is Beautiful’ book are completely unfathomable. It’s a real problem that now anyone can make pretty graphics, the look of the pictures is far more important than their clarity (TV news graphics - yes BBC I am looking at you - are shockingly bad).
And anyone who uses circles (or worse still, spheres) for ‘infographics’ needs to be taken out and lightly shot - it is very hard to assess areas with circles, so using circles of different sizes tends to underestimate the difference between the data.
5 Likes
Isn’t it clear? There’s a legend right there in the corner.
3 Likes
