Infographics and The War on Data

There are few things I hate more in the world than Fox News and Infographics. Although my dislike comes from different things, they both share a common disregard for representing data and in fact, do many things that actively work against the truths that lie in data.

Let's start with Fox News.

I'll just set aside all the misinformation and hate they spew and the entire novel one could right on how the 24  hour news cycle is ruining our society. I'll just focus on this one example featured recently on several blogs :

Read the Blog Here

Here, Fox has distorted the X-Axis of the graph to understate the drop in unemployment. Even a glance, you can see the data is just flat wrong. 9.0% is represented as the same as 8.6%. Over and above that, the visual design obscures the data behind a barrage of beveled metal.

However, if you look at the original chart from the Bureau of Labor and Statistics it isn't much better.

The drop from 9.0 to 8.6 is shown as nothing short of precipitous. 0.4% looks like free fall. That being said, the 0.4% drop is a pretty big drop for a 11 month period historically. I went back and looked at Google's view going back to 1950 and it is a "big" drop. I would argue that looking at unemployment on a yearly basis might not be the best way to evaluate the state of our economy. 

In my high school economics class, I remember reading How to Lie with Statistics. It was eye opening and has made me squint my eyes and tilt my head to the side whenever confronted by a table or chart. The principles of good data visualization outlined in that book and then further expanded upon with Tufte, Few, and McCandless have defined the way I look at design and data. With little effort one can drastically change the way data is interpreted. Whether by accident or with nefarious motives the consequences can be dire. 

That is why infographics piss me off. These ubiquitous towers of pop culture data are used for everything from the Oscars to world hunger to biking, all with equal disregard for the information they are supposedly trying to convey. I don't think anyone will disagree that infographics are the fast food of data visualization. A certain element of fun is needed to capture our attention, right? 

Of course, you could say the same should be true for every piece of data we share from a detailed report to an executive dashboard. Data should resonate with the audience and should be easily understood. It should provide insight and be compelling. I mean, a carefully placed icon or bar chart can do wonders. People just love those little spark lines too.

Infographics veer off in to the absurd when they actually inhibit access to the data. They do this in a couple ways. First, they obscure the data (info, if you will) behind "clever" graphics. Using color or icon as a visual cue is useful but when a giant Wii controller is used to point to a tiny single point of data, you're just filling space.

The opposite is also true. When the infographic is 3000 pixels tall and chocked full of 500 examples of how Facebook is taking off, you haven't edited enough and I don't know what to look at. Look, infographic designers... you're DESIGNERS. This is your job. Communication. You should be using the same principles you use in that Applebee's Fire Grilled Food brochure you're working on. You wouldn't let Mr. Applebee tell you to put every piece of flair on that thing, right? 

Infographics second sin is how they distort data. Just like Fox News, they make mountains out of molehills and just plain misrepresent data. Take the above example from AddThis' 2011 Social Sharing Trends infographic. It's actually one of the best examples of an infographic I have seen, so this is pretty minor on the scale of infographic atrocities. However, in this visualization they do a couple of things wrong:

1. Lack of context—Without any context, like the overall growth per browser, this data really doesn't say anything. I'd be willing to bet that if we overlaid Chrome's user growth, we'd find that there is a strong correlation to the growth of shares. Because who is going to start using Chrome but still share via Firefox? The data is interesting but only meaningful when paired with the user growth. 
2. Too many data points—Do we really need to show Firefox, Safari, and Opera? IE's precipitous decline and Chrome's meteoric rise are the real story. This is a brawl between the two dominant players in the space. Opera is a footnote in history and Firefox is only in the fight because of Fire Bug's dev tools. Strip the nobodies out and focus on the main event. Include Firefox, but only in that  Chrome surpassed it. 
3. No Mobile—Speaking of browsers, where are the mobile browsers? That's a fight I want to see.

Everything else on this infographic is actually really decent. It's simple, readable, and more efficient than just simply reading the information in a paragraph. Most other infographics fail miserably in this area.

Case in point; this one is from Mashable's Top 12 Infographics of 2012 (THE TOP TWELVE!):

This little gem, all 4.9 MILLION browser crushing pixels of it, is simply trying to communicate how airlines are using social media (really going after a controversial topic) but it fails in so many ways.

1. Huge title— If I could load this thing on a mobile device, it would be useless. On my MackBook Air, I have to scroll a whole page just to see data.
2. Same information, 50 different ways of visualizing it—Ed Tufte would be rolling in his grave shouting "small multiples!" if he saw this. And if he were dead.
3. Nothing Compares—No way to compare the data. I wonder how NW Airlines and Jet Blue stack up against each other? 
4. Low quality graphics— If you're gonna make an infographic, really go for it. Don't mess around. 
5. Waaaay too much info— FIVE THOUSAND pixels tall! You lost me at around 800px. Edit down the information just the way you would a good news article. 

I could go on and on. Really. The final criticism, and one that pretty much every infographic suffers from, is that a well written paragraph with a couple editorial images would be a hundred times more effective at conveying the data. 

Here is my challenge:

Start with the data. Edit an unformatted list of metrics down to the bare essentials. Find the correlations and interesting stuff. Then think about the story you want to tell. If graphics are needed to amplify the story, use them. But don't mess with the data. A bar chart is a great way to show comparison. It is your friend. That 3D donut chart? Not your friend.

Basically, apply design principles to information. Something you should be doing anyway. 

Oh, and make it all dynamic with HTML5. This huge JPEG thing is dumb. We can do better.