As has been pointed out frequently lately, there’s a strong groundswell of designers taking it upon themselves to create data visualizations that at first blush appear to be legitimate but upon further review simply reveal nothing more than the designer’s desire to impress you with their ability to create splashy graphics that include data in some way.
Graphic Design vs. Information Design
I recently spoke up on a lengthy Flickr comment thread on an image posted by another designer who is fed up with the noise being produced in the name of infographics, and it’s on this thread that I posited the term “infauxgraphics” to describe the latest spate of work that purports to display data in a meaningful way but eschews the key points of data visualization.
In this thread, a popular opinion by another commenter was expressed that it’s the mingling and conflation of graphic design and information design that’s to blame here. I took offense to this idea that graphic designers are to blame and spent much too long on writing a Flickr comment to explain that frustration. My full comment is excerpted here:
[I'd argue] that it’s a confused generation of designers who believe that graphic design is just the pursuit of beauty that’s the problem. One of art’s pursuits is beauty, but design in any form should always be primarily about solving a communication problem. If the communication method is beautiful, then all the better.
Designers who think all they need to do is make something easy to look at are designers who miss the mark in understanding that design is about making something understandable, and this translates to graphic design, user interface design and information design.
Anyone, designer or otherwise, who wades into information design without knowing the ground rules and when to break them is embarking on a failed expedition, to mix my metaphors a bit.
Graphic design has never been, and never will be about “making things beautiful.” If you believe that, then you have much to learn. The person who made the iPad graphic suffered from a misguided mindset that they can copy the look of something, and recreate its impact.
The popularity of so many useless infographics – the ones lampooned in this Flickr post – is indicative of just how many people feel design is simply the art of making something beautiful.
This is some basic, first-semester design school stuff. The key difference in my mind between graphic design and information design is simply the existence of a data-set that needs to be represented. All of the other tenets of graphic design apply equally to information design. The use of negative space, the use of visual hierarchy, the necessity of choosing the correct color combinations and not least of all, selecting which items to leave out – these are all just as important in information design as they are in graphic design. Visualizing data effectively requires a superset of design skills to be sure, and the best visualizations out there probably also included an entirely separate person acting as a statistician, but that isn’t to say that just because you’re a graphic or interaction designer, you can’t learn information design. Of course you can, you just need to pay attention to what information design really requires of you and not forget the skills you already know. Recognize that just making it pretty isn’t doing your job as a designer to dig deeper into the meaning of the data.
Fetishizing Data Visualization
One of the first problems with designers creating infographics is that a lot of us have started to fetishize data visualization on the whole. We read blogs dedicated to the subject and our pulses increase whenever we see something that contains a lot of criss-crossing lines and circles and (even better) large numbers next to small labels. There’s a visceral reaction here and it is extremely difficult to spend time critiquing something that affects us this way. But critique we must, because there’s a good chance the person who drew this wonderful-looking visualization either willfully ignored – or was blissfully ignorant of – some really solid ground rules when they embarked on their journey of creating the graphic you’re sitting there salivating over.
Beauty is Not Skin Deep
The main dig against the latest round of infauxgraphics is that they make heavy use of big numbers with tiny labels below them. I blame Nick Felton for the proliferation of this (well, ok, not him personally, but people copying his style). I have to admit, I love this treatment and I’ve used it in a lot of places to great effect. But when you’re putting together an all-encompassing piece, make damn sure your giant type treatments aren’t just doing the work of a simple data table while masquerading as more. This is not information design, this is typesetting.
Custom Visualization or Standard Chart?
After having worked in financial information design, I can say first-hand how attractive the idea of creating a completely custom data visualization is over using a “standard,” tried and true chart. The reality however is that standard charts come with a set of well-worn rules and best practices that can help install the constraints you need to create a truly captivating and useful infographic. Diving deeply into each of them is beyond the scope of this meager blog post, but an incomplete list includes bar charts, line charts, histograms, scatterplots (oh man, I love me some scatterplots), box plots, pie charts (ONLY for comparing parts of a whole, and generally a waste of space… we’ll talk about pie charts in a minute). I’m sure I’ve forgotten some, but you get the point.
The main idea here is to study these and their best practices, chances are you can use one right off the bat, or can modify one to meet your needs fairly easily. Unless you have data that requires a completely new way of thinking about things (or you’re Ben Fry or Golan Levin), you’re gonna be able to find a visualization that’s reaaally close to fitting your data set.
Look at the silly pie chart to the right. This made the rounds a few days ago, primarily because it’s the first set of iPad and iPad app data to become available, and it was visualized in a way that on the surface didn’t look half bad. But the designer took extreme liberty with this pie chart, because you are intended to view the entire circle as the median iPad app purchase price ($4.99) and compare the median iPhone app purchase price ($1.99) to the whole circle. Comparing the size of two values is a bar chart’s job, not a pie chart’s, and these two values in no way, shape or form add up to a whole or 100% of something.
Seriously, stay the fuck away from pie charts. And for the love of god, if you have to use one, use only one. I’m guilty of this very transgression and I’m in the process of switching all my pie charts to something else because pie charts are really, really, really useless, almost all of the time. The main things people like about pie charts are:
- They are circles, and everyone loves a circle. Circular things look like important things. This is not enough of a reason to use one.
- They can contain artfully segmented pieces inside of them, and designers like to break things up into segments. This is a valuable skill, but it’s being misused when it’s a pie chart.
- We like to shove too many things into a pie chart. If you have more than about 5 slices, use something else. If the things in the pie chart are not part of a whole that equal 100%, do not use a pie chart.
- We like the look of a lot of pie charts on one page together, but one pie chart is completely incomparable to another one. Humans just can’t perceive the differences between circles, whether they’re sized differently or include differently segmented pies, we simply can’t detect the differences between similar items very well. Bars work so much better.
I’m not really sure how to conclude this other than to recap that just because you’re a graphic or interaction designer doesn’t mean you can’t also be an information designer, it’s just that you shouldn’t fetishize the data visualization. Learn the ropes first, because the good things about good visualization can’t be copied at a superficial level just because you think they look neat, there needs to be a purpose behind every decision you make, just like when you’re doing your other design work. Don’t throw out your graphic design expertise either though, as it’s important to create something people want to actually look at. Just make sure you’re not making something with the primary goal of making people impressed with your skills. Those things fail when put under the microscope of having to understand what the data is trying to say.