I'm currently sitting in the airport after attending the Joint Statistical Meetings (JSM). It was actually my first JSM; as my PhD starts finishing up I figured I should attend at least once. I'm glad I went although I doubt I'll go again.

The conference was located in Vancouver which was delightful. My only complaint with the city was the accumulation of Canadian coins (but not pennies!). There were heat advisories while we were there with temperatures rising to a sweltering 28 degrees Celsius. Although I do feel it got quite a bit higher in our airflow-challenged AirBnB.

The talks were decent although I'm generally less than enthused about attending lectures in person. I'd much rather they be filmed so I can watch at 1.5 speed and pause/resume. A few highlights:

  • "Theory vs Practice" in which a panel of some big names in stats discussed the current state of play with respect to statistical theory and statistical practice. While things are potentially changing, the consensus was that an academic statistician without theorems will have a hard time. Made me quite glad I'm going into industry.
  • My Google supervisor Dennis Sun mentioned my internship project Meterstick in his talk.
  • Chris Genovese plugged emacs and org-mode in his talk on reproducability.
  • I saw Hadley Wickham edit his book live during his talk on reproducability. He also showed a R script for automating managing a shared electrical bill. I'm inspired to automate my own menial tasks.
  • Cynthia Rudin on interpretable machine learning. Showed the NN tool This looks like that

For my part I gave a poster on my ABC work during the opening mixer. There was a surprising number of people who came by. It was a quite strange experience trying to explain mostly my entire thesis on a single poster. It's a bit reassuring to explain my research to non-statisticians and realize how much I've learned. Now doing so with statisticians feels even better.

Of course, JSM isn't all about, or even mostly about, the talks. The main benefit was socializing. As not a particularly social person, I felt I didn't get the full advantage of this aspect of the conference. Most of my time was spent with people from my cohort: we could have done that in Pittsburgh. But it was fun exploring a new city with my friends and meeting the occasional new person.

If there was one thing I'd do differently it would be to sign up for the Career Service. While I wasn't 100% excited for many of the companies at JSM, it would have been nice to get a big dose of interview practice before starting my job search. And who knows, maybe I'd end up with an offer in the process.

All in all it was a good experience. I'd recommend any grad student attend at least once. Especially if it's in a good location and you have funding. Speaking of which, it is in Denver next year…