My Favorite Things

Raindrops on roses and whiskers on kittens
Matrix transposes and not overfittin’
Well-maintained packages, clear docustrings
Networks with very low latency pings…

Speed that impresses from large L2 caches
Deepfakes detected with watermarked hashes
Lightweight code linters that fix review dings
Matroids and monoids and fields and rings…

Here is a list of my favorite academic things: the experiences that keep me doing what I do. These aren’t professional successes - I don’t get that happy or excited when papers are accepted, for example. They aren’t even intellectual experiences, really. Rather, they’re visceral, emotional experiences. My favorite food critic has a passage about the visceral joys of cooking, like slicing a ripe melon in half. Cooks work with food, and potters with clay, and I work with data, and here are ten things I like about doing that:

  1. Getting new data - especially if I’ve waited a long time for it. Sometimes it’s just a link on a webpage. Sometimes I’m running some scraping script for months and the data accumulates like snow - I always find that very comforting. Once the data came on a 12 terabyte hard drive that I carried across the country very carefully, in a lockbox I borrowed from my boyfriend’s parents.
  2. Fast computers. I get a thrill from working with them, the same way some people get a thrill from fast cars. I like the moment when I’ve rewritten my code to run on hundreds of cores and I fire up a cluster that costs more money than I’ve made in my whole life and I can feel it kick into gear and watch in htop as the processes rise and fall like red-green flames.
  3. Students who teach me things. It makes my day when I review some clever analysis a student has done and am like – not only would I not think to do what you just did, I wouldn’t even know how to do what you just did. Relatedly, I like when a student has some promising but preliminary result and I have to wait til next week to see what they’ll find next. It’s like waiting for the next Harry Potter book.
  4. Feeling like my work has helped someone.
  5. The magic of some machine learning/statistical techniques. It still blows my mind when I download a neural net and it correctly classifies some random image I plucked off Google. (I don’t really even like deep learning because it feels like black magic - but it’s still pretty magical). Bayesian inference is another example: I always feel better when I can write down a Bayesian model of the world, and fitting it in Stan is pure magic.
  6. Finishing my last meeting of the day and knowing I have the rest of the day to myself to think. I’m pretty antisocial, so if I’m really lucky, this happens at 10 AM :)
  7. Working with long-time collaborators who have become friends. You trust them enough to know that even if you drop dead, they’ll get the paper done (though you might not care much at that point).
  8. Those oh-so-rare moments when I generate a striking new result, and I know that I’m the only person in the world who knows this new thing.
  9. College campuses - I’m not sure I’ve ever found a campus I don’t like. Stanford is my first love and for my money it’s the most beautiful school in the world; don’t @ me. But I even like Stanford’s lesser rivals. I like walking through the pine trees on the Berkeley campus and past the fountains at USC; the colleges of Oxford and Yale, how every time you pass through a new archway you emerge into a courtyard too lovely to be real; how you can stand in the middle of Columbia campus and forget you’re in New York City; the Harvard and UChicago quads on summer evenings, with the light slanting golden through the trees. I even like - bless its heart - MIT, even though they’re too Busy Doing Science™ to construct buildings with any aesthetic appeal or, for that matter, names instead of numbers. When I go to Boston, I always try to visit the corner between the Stata Center and the Broad Institute just to revel in the density of famous scientific buildings.
  10. Female and non-binary friends and collaborators in a male-dominated world. Meeting younger female students and thinking - wow, you’re going to change this field for the better. We best make sure you stick around.
Written on November 13, 2020