Comments on: Political Pie Charts

By: Jon Peltier

Jon Peltier — Wed, 22 Jul 2009 04:46:12 +0000

Thanks for the link. I've been reading up on color vision, and how the human three-receptor system gives rise to what we think of as color vision. Absence or deficiency in one or another type of receptor gives rise to the various color vision deficiencies. Most placental mammals have only two receptor types, leading to reduced color discrimination. It is thought that a mutation in early primates split one of the receptor types into two types, giving primates improved color vision. This is nothing compared to birds and reptiles, which have four or five types of receptors (depending on species), extending into ultraviolet. We can't imagine the richness of colors in their world. See Color Vision in Birds for a nice overview, or Color Vision of Birds for the gory details.

By: Jon Peltier

Jon Peltier — Wed, 22 Jul 2009 02:04:54 +0000

I just remembered an interesting story about a college friend. He had a part time job for the phone company, connecting the wiring for phone lines coming into the dorms and going up to the rooms. Well, Wayne was colorblind, so he was unable to see the color coding on the small wires. Much of his work didn’t require this, but when he was hooking up the specific wires fpr specific phone lines, he’d have to ask one of us to lend a hand (or an eye).

By: Naomi B. Robbins

Naomi B. Robbins — Wed, 22 Jul 2009 02:00:55 +0000

Here’s another resource on color. It’s a master’s thesis written under Cindy Brewer.

From http://www.personal.psu.edu/cab38/ColorBrewer/ColorBrewer_updates.html

“Steve Gardner’s Master’s thesis evaluated the ColorBrewer schemes for colorblind readers. I’ve linked his whole thesis with his permission (6.1MB .PDF file or 5.5MB zipped which isn’t smaller but it may download more smoothly). He has recommendations that I have not yet incorporated into the schemes, so I want to get that info out to you. Please cite his work if you find it useful. Also, I hope his work prompts further student research. Steve is now working at the National Geographic Society.”

By: Jon Peltier

Jon Peltier — Wed, 22 Jul 2009 01:35:31 +0000

Tim -

Thanks for your comment. I recall seeing your post some time back, and thinking that, hey, Microsoft DID do something right with all those enhancements to Excel 2007’s colors. Then I promptly forgot about it.

The ColorBrewer utility I’ve blogged about doesn’t correct for color vision deficiencies, but it at least indicates whether a selected palette is good, iffy, or bad for someone who has difficulty distinguishing colors.

By: Tim Wilson

Tim Wilson — Tue, 21 Jul 2009 23:16:00 +0000

Hah! I read the post and had the same thought Naomi had and went and did some poking around with http://colorfilter.wickline.org/ before I scanned the comments. I checked the most common form of color-blindness, deuteranomaly, which is the most common form (affects ~4.9% of men) and the charts work fine. Deuteranopia affects ~1.1%, and there is an issue there, although I was still able to distinguish between the two wedges in question.

BUT, the larger point…is that color blindness (and “printing on a black-and-white printer” and “photocopying”) is another reason that a bar graph often trumps a pie chart — what takes multiple colors (or horrid patterns, or evil color palettes) in a pie chart can often be done monochromatically with a bar graph.

I actually hope to get around to developing an Excel color palette that is both aesthetically pleasing to people without any color deficiency *and* that works for those with the most common forms of color blindness (there are, I think, four different forms of color blindness that occur in roughly 1% of men or more). I did some experimenting with Excel 2007 a while back to at least check their default palette: http://www.gilliganondata.com/index.php/2009/06/18/data-visualization-that-is-color-blind-friendly-excel-2007/

As for fivethirtyeight.com, some of their sidebar charts during the election actually confused me — lots of information crisply presented, but I often found myself wishing for an easy link to an explanation of how to interpret them.

By: Jon Peltier

Jon Peltier — Tue, 21 Jul 2009 00:02:59 +0000

I have thought much of what 538 has posted is well done, especially around last fall’s election, but occasionally they miss. The climate change chart wasn’t terrible; they could have used more care with labeling and a lighter touch with the lines.

Thanks to Naomi and VisCheck, looks like I will have to change my default palette again.

By: Chris Read

Chris Read — Mon, 20 Jul 2009 20:46:18 +0000

Jon,

Just popped onto FiveThirtyEight’s website to see what they’re all about. It seems they have employed someone who’s just been shown how to start Excel, given their shockingly novice charting skills. Here’s another to cringe at:

http://www.fivethirtyeight.com/2009/07/challenge-to-climate-change-skeptics.html

Among the many faults, you’ve got to love a 90 degree head tilt to read x-axis labels…. I look like my West Highland Terrier when he’s trying to work out what I’m saying.

Very interesting comments on VisCheck, I shall definitely be keeping this in mind in future applications!

Chris

By: Naomi B. Robbins

Naomi B. Robbins — Mon, 20 Jul 2009 14:14:45 +0000

I just want to make it clear that I am not recommending the original Excel default colors. There are many attractive color schemes that all readers can interpret. I also want to take back my comment that the protanope and tritanope were similar to the deuteranope for this example. Clearly, from Jon’s figures above, we notice that the colors they see they are distinguishable.

By: Jon Peltier

Jon Peltier — Mon, 20 Jul 2009 13:14:38 +0000

Naomi - Thank you for reminding me about color vision deficiencies. I've reproduced your experiment here using the image tester at VisCheck. The Excel default colors are to the left, my colors to the right.

Original Colors

Deuteranope (a form of red/green color deficit)

Protanope (another form of red/green color deficit)

Tritanope (a blue/yellow deficit- very rare)

By: Naomi B. Robbins

Naomi B. Robbins — Mon, 20 Jul 2009 12:31:39 +0000

This comment only addresses the color choice where Jon says “using arguably better colors.” I placed the original Excel colors and Jon’s colors in vischeck (www.vischeck.com) to see how vischeck says a person with color vision deficiencies (politically correct term for colorblind) would see them. I’m showing how a deuteranope (one form of color vision deficiency) would see them but the other forms are similar. Note that a reader with color problems cannot distinguish the red and green of PeltierPie2 bit can with ExcelPie2.