Data visualization is a fascinating area. I also think it’s just begging for innovation.
Analytical types usually aren’t very good graphical artists (guilty). Their palette is defined by a few clunky buttons on Microsoft Excel’s ribbon, or, if they’re really eccentric, Apple’s Numbers.
I’d like to see more graphics experts work with number crunchers to find creative and accurate ways of displaying complex data.
Some of the most effective modern data visualizations I’ve seen make good use of scale and employ everyday objects. They make numbers relatable.
For example, this visualization of US debt does a great job of conveying the relative size of very large numbers.
The human mind doesn’t naturally resolve very large quantities (millions, billions, trillions,…) because it wasn’t very important from an evolutionary standpoint.
A hundred thousand years ago, it was important for the tribe to understand the difference between having one apple versus fifteen apples. Cavemen never dreamed their descendants would have to weigh one trillion versus fifteen trillion of something.
In a world with increasingly short attention spans and increasingly large numerical problems, we need people who can portray complexity in simple, easy-to-understand pictures.
It’s time for a renaissance of the cave drawer.
{ 2 comments… read them below or add one }
Many earlier societies had the most rudimentary counting. One bison, two bison many bison. They couldn’t fathom the larger numbers. Today when people quote large numbers, 50 million starving in Africa, 40,000 ded on a single day at The Somme, 5 billion tonnes of CO2 locked up in forests I find it equally as meaningless. What does even a pile of one million dollar bills look like, would it fit it a briefcase or a wheel barrow?
Depending on your interests there are many good books and tutorials in this area. I teach a visualisation seminar for students primarily from applied math and stats but also some more numerate ones from criminology, economics, computational linguistics and cognitive science. We focus on how to choose what aspects of what data to present and then how.
Depending on your background there are books using R, matlab, processing, python and so forth. These books are all fairly realistic in assuming that no graph or figure should be your life’s work.
I am guessing R might be your friend, but I can make an informed suggestion with a little more information.
Cheers, Jif.