Sankey, Connecticut

You've seen Sankey diagrams before - if you're a fan of visual communications.  (Minard. Napoleon. Tufte.)  They're considered a "flow diagram" - the volume of flows is represented by volume (width) of connections, frequently color-coded for distinction.

Yeah, yeah, talk talk.  Check this out.  From a state publication on waste streams, no less!

Two cool points.  The big point doesn't require explanation - namely, that this doesn't require explanation. You can read it.  It makes sense.  The smaller point - ask yourself, did you really read the text at the bottom, or did you glance over it and then flick back up to the visual?  Yeah, thought so. Me too.  

Written documents are wonderful, but (I'd argue) they're not the best tool for discussion of complex issues, particularly complex issues involving quantities and dynamic relationships.  Imagine a committee arguing with the support of a dynamic Sankey diagram, visible to everyone.  Better use of their collective smarts?  Better outcomes?  

Please, please call us with a Sankey-esque problem, even if you're not from Connecticut.

Sankey Linky: http://www.ct.gov/deep/lib/deep/waste_management_and_disposal/solid_waste/transforming_matls_mgmt/summit_1/handout_unlocking_the_value.pdf

Deft Animation From The FTC...?

It's fun, humbling and inspiring to see good explanation media in unexpected places.  The Federal Trade Commission has quite a few media resources at http://www.consumer.ftc.gov/media.  This one was particularly fun.

The combination of hand-drawn visuals with multi-plane animation, object movement and camera movement - that's not easy, kids.  Notice the reflections on the hood when the car is moving, and the birds flying alongside about 12 seconds in.   And the camera zooming in, as the car zooms up to the stop sign at @25 seconds.  As I said, deft. Whoever did this - nice, thanks!

-md

Source: http://www.consumer.ftc.gov/media

Big Data. Little Pictures.

I had the opportunity to speak about content and understanding at the Explainathon, hosted & mosted by Adam Zais of Wistia, on May 22. (Yeah, my birthday.)  No doubt the video of the talk will end up here.

One remark along the way - "Why is it that the answer to 'Big Data' is 'Little Pictures" prompted a followup question via Twitter.  My response was 'because screens are so small right now.'  It was a wisecrack, but I think it's a serious (and darn interesting) topic. 

Big data is one of the unexpected consequences of the digital shift.  Let me put it this way: you viewing this blog post generated at least a hundred data points - most of them probably invisible to you.  

  • We use a brilliant tool called Woopra to see web-site visits in real time.  Your IP, location, and everything else your browser advertises about itself, are captured there. (Yeah, you with the huge monitor - we know the resolution!)
  • We run Google Analytics.  All that stuff is trapped there as well.
  • Google itself grabbed a bunch of data if you got here via search. Bing, ditto.
  • Our web host probably logs most of the same stuff.

And that's just scratching the surface.

So I'm supposed to find the meaning in all of that (multiplied by thousands, across months) by staring at a flat screen?  Little pictures are the best we can do - for now.

I expect to see visualization frontiers expanded rapidly in the next @7-10 years. We'll start cost-justifying solutions that enable us to see more, based on the value of pattern-and-meaning to be had from all that big data.  There's an optimistic and humanistic statement mixed in there.  I think the job of finding (ascribing?) meaning will be human-first machine-second, for quite a while, and I think it'll be primarily a visual job.  Once we know what we're looking for ("look" - a visual reference...), we'll tell the machinery to go look for more stuff like that.

Or maybe this is just a big rationalization for those 29" Dell monitors that would fit SO nicely on my desk.