Why Beautiful But Dumb?

Good information architecture balances a lot of competing elements: design, business needs, marketing needs, stuff the CEO wants, things the tech department wants to try out, and user needs. Without IA ,the web, and the world, tends to start drifting towards what looks cool (the beautiful part) in favor of being smart (hence the dumb part).

Monday, June 20, 2011

Information Visualization: The Good, The Bad and the Nauseating

We are hard at work on an information visualization project over here, so we've been exploring the #infoviz world and all its beauty and, well, idiocy.  As much as there is room for mis-steps along the information architecture path on a traditional site, when it comes to data visualization it seems that the opportunities to get it wrong grow exponentially.

This is my favorite disaster thus far:

I've been spending a lot of time looking at this picture trying to understand exactly what went wrong so I don't make the same mistakes, and here's what I've come up with:

  1. The title is bizarre (and actually a little revolting once you figure out that the contents of the jars is supposed to be blood), which means you have to read the fine print to even begin to understand what you are looking at.
  2. You can't just scan the picture and immediately determine which conflict caused the largest number of deaths, because in order to do so you have to make a snap judgement about how much liquid different vessels can hold, and this is simply too much work.  Does the mixer hold more than the pitcher?  I have no idea, and I am someone who does a lot of cooking.  This means that in order to compare conflicts you have to read the accompanying text.  I bet if someone did an eye-tracking study of this image the points would be all over the place as users attempt to make sense of what they are looking at, and compare points.
  3. There are two key data points that most information visualizations need to get across - comparison of parts and the whole.  Users need to understand how the smaller parts compare to each other, and then they also need to be able to step back and see the information as a whole picture.  This image does neither of those well.
  4. When it came to the whole picture concept, the designers got confused about what point they were really trying to get across.  I'm assuming that the big idea for this originally started with a sentence like, "Hey, let's show how war is bad," followed by "ooh, let's use a kitchen metaphor."  But the designers got so entranced by the kitchen metaphor that they ended up muddling the "war is bad" theme.
  5. The most important point: the metaphor the designers used is, frankly, appalling, in addition to making no sense.  What does war have to do with cooking?  And also, I don't want to look at blood in kitchen appliances.  The overall metaphor is so revolting, and therefore distracting, that it overwhelms the original message the designers were attempting to communicate.
New York Times Book Review

I have also learned a lot about information visualization from the infographics in the New York Times Book Review.  I'm not entirely sure why the book review section is so entranced by infographics, but they frequently use them to convey information about book sales.  They never do them well and never convey any interesting information.  Here's the one from this weekend (apologies for the image quality - my scanner is on the fritz):

What are we supposed to be learning from this infographic?  Here are my guesses as to what the editors of the Book Review think they are telling us:

  • Eric Larson sells a lot of books, in four-year intervals
  • These books are usually about death and murder.
  • Some books do better than others?
  • The books start out selling a lot of copies, and then drop off?
Unfortunately, they're not conveying any of this well, and moreover, a lot of it isn't surprising (and therefore interesting).  If they really are trying to show the time intervals of his Larson's book releases then they should have used a graph that would have made those intervals more obvious.  IF they are trying to show that books start out selling well and then drop off, then I ask: why is this interesting?  Don't most books do that?   And if they're trying to convey that his books are usually about death and murder, that seems better communicated in an essay or a little blurb than in chart format.

But what really confuses me about this graph is the bottom area where they are (I think?) contrasting weeks on the list with the highest rank the book achieved.  What do these two numbers have to do with each other?  And why is contrasting them interesting?  I don't know, and probably no one else does either.

So here's what I've learned from the book review infographics:
  1. Do not create infographics just so your designer has something to work on.  Do not create infographics just because you have some stuff that includes numbers.  You should only convey information in chart form when the numbers tell an interesting story, and when those numbers can also be abstracted to tell a larger story.
  2. Do not compare numbers that have nothing to do with each other.
  3. Understand what the key piece of information you are trying to communicate is before you start making the graphic.
Create Your Better Life Index

Create your better life index - OECD Better LIfe Initiative
One infographic that does a good job handling a lot of data is the create your better life index at OECD Better Life Initiative.  There is a clear connection between the data users check off on the right side of the screen and the chart.  In addition, the overall picture tells an interesting story (different countries make people happier depending upon what criteria are important to you).  It's also nice that you can display countries both alphabetically and by rank.  I do have a few quibbles with the chart:
  • The numbers over on the left don't really mean anything to me.
  • The level of detail in the scale on the right is probably overkill.  Do I really know the difference between Jobs being two circles of importance to me versus four circles of importance?  Wouldn't it have been simpler for users to just select if jobs are important or not?
  • The "by rank" display option is much more informative than the "alphabetically" option, so it might have been a better choice to set "by rank" as the default.
That said, the infographic is still fun to play around with, conveys interesting information, and allows users to view both details and abstracted data.