Edward Tufte and the Big Picture
I saw Edward Tufte speak yesterday. He led the day off showing a data visualization similar to the one I linked to above. The rest of this post has nothing to do with this, but it was really amazing, and I wanted to share! As an aside, this may be the way I might finally be able to embrace classical music. I am such a visual person that this way of "showing" the music really made a huge difference in how I interacted with it.
Tufte covered a lot of ground during his talk, including his legendary loathing of PowerPoint. But the most striking take-away for me was the concept of finding one data-rich graphic that can be used to communicate many many points of data. He gave the example of a fire chief who has to report to the city council on the fire department's performance.
What if the fire chief were to take an aerial map of the city, and places a red dot on each building where there was a fire ("I'll even allow a flame icon for really big fires"), a yellow dot for each false alarm, a green dot for each passed inspection, a purple dot for each failed inspection, etc.? Then the chief would present that on a big piece of paper, 11" x 17". (Tufte says this is still the best way to present high-resolution data, but in a few years, when we all have iPad-like tablets, that will change.)
Each city council member would engage differently with the fire map. Some would zoom immediately to their own districts, others might look at the discrepancy between where the fires happened and where the inspections took place. This one data visualization meets the needs of multiple recipients. In this hypothetical example, the chief would present this piece of paper, then give everyone time to read/interact with it. Then the chief would point out some interesting highlights, followed by questions from the council.
This is brilliant. And I've been considering what that one big data graphic would be for fundraising. Would it show a complex view of the giving pyramid, with prospects being fed in by prospect research and the annual fund? Would it be multiple years of giving data? I've been chewing this over and am really intrigued by the possibilities.