Like many geeks I secretly covet Tufte’s serenely assured design sense, and fetishize his pronouncements. That said, most sparkline implementations I’ve seen have left me confused. The sparklines that Nat is generating of Netflix rental data are different. They’re great. Great in the sense that they’re small, and trivial, communicate exactly what they’re supposed to do.

Looking at my last 16 months of rental history , you can see that I almost always spend on an average more per movie then I would have at the corner video store ($3/video). Either I’m a sucker, or am I getting some other value out of it. Either way, that itty bitty graph packs an amazing amount of info in.

And he’s got a service setup where you can generate your own.