|Small vocational joys
||[Oct. 30th, 2012|12:31 pm]
Two work things what are small but awesome:|
1) Yesterday I made a complicated plot of data (which incidentally looks kind of like a toddler drew a bunch of lumpy circles on a map of North America).
Me: [launches into long explanation of what exactly we're looking at, ending with...] so what you can see with this figure is, in this region, [thing 1] is noticeably bigger than [thing 2], whereas in this other region, it's the other way around. My question is, do you agree with that interpretation, and is that result at all interesting?
Colleague: [thinks a while] ... Yes. And yes.
2) Today, I tracked down this rule of thumb for how many bins to divide things up into, which I at one point found and wrote down but neglected to write down where it came from. First, I discovered that's it's actually a well-known formula, which is nice in terms of justification. And then in reading up on it, I found out that it has an implicit assumption that only hold for half of what I'm doing. But I also found an extension of it that tries to improve performance when that assumption doesn't hold, plugged some typical numbers for my work into that, and got...
Exactly the same number I was using already! YES! \m/