The problem is that the idea of visually representing information has caught on with the general masses -- and 90% of everything is crap.
It wasn't really all that long ago that terms like "infovis" or "datavis" were novel and new to the great majority of people. Sure everybody was familiar with standbys like bar charts and scatter charts, but the idea that you could cram complex multi-dimensional data into a single, novel, picture was relatively new as a discipline.
So now everybody and their aunt wants to "visually represent information" even if it doesn't make sense to do so, or the representations chosen artificially distort the information.
The problems are many but boil down to three essentials:
- the art department produces this stuff ("it's visual, so it must be the art department's job"). This is a department that is fundamentally wrong for producing representations of quantitative information. But I'm not sure giving it to the bean counters, statisticians, engineers or other nerdly disciplines is any more right.
- it's hard to come up with new and novel ways of representing information. Tufte's come up with approximately one - sparklines. And he's widely considered the king of the discipline. A couple others get bandied about ad-naseum, Napoleon's March for example, and as a discipline it seems more geared towards taking apart other people's efforts at novel visualizations rather than producing new, good ones. Innovation in Infovis is rare.
- even when falling back on traditional techniques (bar charts everywhere!), they are open to abuse or misinterpretations depending on how the data is viewed. Have 1000 entries but only room for 10 in the graphic? Drop the other 990! Use logarithmic representations when linear would be better. Focus attention where you want with colors or other indicators. By choosing how the viewer gets access to the tables full of boring numbers, you can influence what you want them to see.
Infographics can be an entertaining way of educating the non-quantitatively minded reader without presenting tables full of boring figures. But I agree they are widely misused.
I go one further to say that infographics, as referring to the kinds of things that should be produced, should almost always supersede tables of boring figures. As it turns out, even if you're quantitatively minded and interesting in the particular topic being analyzed... it's still better to see a well-designed chart.
I mean, I can't help but look at the chartjunk tsunami and see opportunity instead of just so many dire straights.
---
Also I don't feel that the real need is in innovation of new standard charts but instead in the field of merging visuals with quantities in a way that is fluent to human psych.
Having worked in Academia, charts are often used to reveal trends in large data sets. This is particularly true these days when the data is gigabytes of multi-dimensional measurements that require processing to derive the underlying properties. The only good way to present data like this is in charts.
Graphics are now very important to presenting your science! The results can't speak for themselves, because no one can understand them in less than a few days of study, so they need a translator.
Sometimes people do create the equivalent of link-bait info-graphics that do their best to make a small result look large. It's usually balanced by the rest of the paper being light on details.
These infographics are just the newest version of sound-bytes and headlines. They don't tell the whole story, but they tell enough that you can tell your friends and pass out the link, and feel like you know the subject without having to spend more than 5 minutes reading. What we're missing is the balance of listing methodology and explicitly listing sources. There's little to no onus on the publisher to fact-check their data.
As long as we live in a society with 1-minute clips on the evening news, we'll have crappy infographics.
It wasn't really all that long ago that terms like "infovis" or "datavis" were novel and new to the great majority of people. Sure everybody was familiar with standbys like bar charts and scatter charts, but the idea that you could cram complex multi-dimensional data into a single, novel, picture was relatively new as a discipline.
So now everybody and their aunt wants to "visually represent information" even if it doesn't make sense to do so, or the representations chosen artificially distort the information.
The problems are many but boil down to three essentials:
- the art department produces this stuff ("it's visual, so it must be the art department's job"). This is a department that is fundamentally wrong for producing representations of quantitative information. But I'm not sure giving it to the bean counters, statisticians, engineers or other nerdly disciplines is any more right.
- it's hard to come up with new and novel ways of representing information. Tufte's come up with approximately one - sparklines. And he's widely considered the king of the discipline. A couple others get bandied about ad-naseum, Napoleon's March for example, and as a discipline it seems more geared towards taking apart other people's efforts at novel visualizations rather than producing new, good ones. Innovation in Infovis is rare.
- even when falling back on traditional techniques (bar charts everywhere!), they are open to abuse or misinterpretations depending on how the data is viewed. Have 1000 entries but only room for 10 in the graphic? Drop the other 990! Use logarithmic representations when linear would be better. Focus attention where you want with colors or other indicators. By choosing how the viewer gets access to the tables full of boring numbers, you can influence what you want them to see.
Infographics can be an entertaining way of educating the non-quantitatively minded reader without presenting tables full of boring figures. But I agree they are widely misused.