Category Archives: Printed Information

Minard for Sale

Daniel Crouch Rare Books has a catalog of various data visualization offerings including a nice  collection of Minard works for a hefty £400,000. Some of the works I was unfamiliar with and yes, a copy of the Napoleon map is one of them. You can purchase the catalog from their site and download a PDF here.


A Redesigned Oscar Winner Card

Redditor ShinyTile points out that poor graphic design may have contributed to Sunday night’s Oscar mixup.

I agree and took 5 minutes to redesign the card.

As ShinyTile points out, the Oscars logo catches the eye first, and in this context is entirely irrelevant to the purpose and usage of the card. I assume the cards are nice keepsakes (in addition to the statues), and so I’m okay with keeping the logo, but minimizing it and making it the last thing the eye might read. In its place at the top center, I would place the category in the same Oscar logo gold. That should be the first place the reader’s eye goes and it should serve to confirm the category winner about to be announced. But immediately after the category is processed by the reader, the next thing is the winner and the first thing announced—big, bold and in all caps.

I’m okay with the title being all caps, but I would make the additional information (in this case the producer names), sentence cap as I think this is easier for the eye to read, especially with longer and more complicated names. The only things read aloud are in black and the other two items are in the less prominent gold.

Just a suggestion…


Also, the LA Times points out that the mixup could also partially be due to poor envelope design.


Consumer Reports and Y-Axis Shenanigans

There are many data visualization sins, but the one that makes my blood boil is manipulating the Y-axis. Fox News is the master at this, so often just deleting the Y-axis entirely to invent the data story that syncs with the network’s political narratives.

While Fox News generally is creating a story that isn’t there, usually when I see Y-axis shenanigans, it is to exaggerate a legitimate story. And generally, the story is a good one to begin with that doesn’t need exaggeration.

I was not been a big fan of the Consumer Reports redesign from a graphic design perspective, but it has slowly gotten a bit better. But the really need a stronger editor (and backbone) when it comes to data visualization. Your data tells a good story to begin with—there is no need to visually lie to your readers, as they did recently with these charts showing secondary market ticket prices. Because the length of the bars indicate value, distorting them is essentially telling an untruth to your audiences. (No, $22.58 is not one quarter as much as $36…)


Just being visually truthful doesn’t make your story less impactful. And if it does, then you need to get a better story! This would be a far more truthful design:



Or, since it is a trend, I would probably suggest a line chart:



Different Audiences, Different Process Graphics


I came across a simple, but excellent example of process graphics in Michael Bierut’s new book, How To. Above are two pages from the printed brand guidelines for the design of United Airline’s once low cost sub-brand called TED.

As part of the design and branding work, Michael’s firm Pentagram wanted to show how the different divisions of United Airlines all fit together. Instead of a single visual (“process graphic”), they created two: one for internal audiences (on the left) and one for external audiences (on the right.)

This is a great reminder for me that one story does not always mean one single graphic. Think about your audience, think about their level of knowledge, think about what you want them to take away from your graphic.

Below is the full page from Michael’s new book.



An Interesting Chart from The Wall Street Journal

I thought the above chart on the left was interesting, if not quite as effective as another chart might be. Most would have jumped to a pie chart, or maybe a stacked bar chart, but the WSJ today created a…”Dot Bar Graph”? I don’t know what you’d call it, but it did take me a moment to figure it out as a reader. And then after I did figure it out, second guessed myself due to the removal of the %’s after 40%. (Were there only 27 women polled, I wondered?) 

After a few moments, I realized that all the numbers (which are percentages) add up to 100%—something that is not so intuitive since the scale ends at 40%.

Perhaps they didn’t want to do a bar graph (which is perfectly acceptable when all values add up to 100%) so as not to confuse things with the bar graph on the right?

So, in the end, it’s “interesting” to me. But maybe a stacked bar chart or tree map would have been better?

visual training presentation