Data visualization

Today, one new link about data visualization, and one older one which I hadn’t blogged about yet.

The new one first: Here is a video of home prices in the US from 1890 to the present, visualized as a roller coaster ride. It’s a cute idea. I would have loved to see more data and more context (especially of the era the coaster car is in). But it was made using Atari’s Rollercoaster Tycoon 3, so it’s really a testament to using cheap, unconventional methods to show something.

For a much more mind-blowing look into what can be done with data visualization, have a look at Hans Rosling’s talk at TED 2006. During his talk, he uses very cool data visualization techniques to draw counter-intuitive conclusions from data on human development. It is very convincing, and a powerful talk.

I find this topic interesting for several reasons.

First of all, data visualization, as a part of interface design, is something that can be used in games to clearly show important aspects of complex data, which is what you need in some genres. Simulations, tycoon games, city-building games and 4X games all seem obvious candidates for this, although I can also see it used in hard-core analyses of less obviously complex games (I say this as someone who analyzes World of Warcraft combat logs for raids with up to 40 people for fun). Data visualization techniques might be a powerful tool to make complex simulations more accessible. Inversely, it should also be possible to use techniques from games and game development to improve data visualization applications.

Secondly, the game development process itself can generate complex data, and data visualization could be used to analyze it. Bug reports and repository access statistics come to mind, although right now I can’t think of anything I might want to tease out of that data that isn’t fairly obvious and can be extracted using trivial methods. But you never know. A more promising field could be the analysis of play-test sessions, one of my pet subjects. Having data visualization techniques (not to mention knowledge of statistics and other mathematical fields) in one’s toolkit should be of great help. (And this seems as good a time as any to link to Edward Tufte.)

Thirdly, it might be interesting to build (serious) games around data visualization. I am currently imagining a variant of Clue… “The drop in child mortality in Paraguay in the 60s is responsible for the increased GDP in the Mercosur region in the 90s… with a candle-stick in the library.” Hmmm. Needs work.

Finally, there are areas where games are superior to data visualization (which, after all, uses static data) in conveying deep insights. This is a fairly fundamental idea that is not often discussed (I can’t think of many people apart from Chris Crawford who have written about it directly), but interactivity is a way of conveying information just like text or images, and there are types of information, such as causal relationships and complex processes, that are better conveyed using interactivity than anything else. So, in theory, there could be a game developer at TED one day showing how games can produce powerful new insights into the world we live in.

(Roller coaster link via Jason Kottke, Hans Rosling link via all over.)