I went to PyData Boston in July. The videos are just coming out now, so here are a couple of things that I enjoyed.

Travis Oliphant, the CEO of Continuum Analytics & author of NumPy gave the opening keynote. It was a great overview of what Continuum Analytics is up to (basically making wonderful software like Anaconda and Wakari and a million other amazing-looking things.).

When talking about Numba, he said “In my mind, there’s no reason to write C++ anymore” for scientific computing – you can just use Numba and decorate your Python functions with @autojit.

Goodness.

Lynn Cherny gave a great talk about detecting sex scenes in Fifty Shades of Grey. She talks about using Mechanical Turk to build a classifier. It’s super interesting and hilarious. I’m really happy that I know that she exists now. I also found out that she maintains a mailing list for data visualization jobs. Here’s the video.

She said that she does all her data analysis in Python, and then writes frontends in Javascript (using D3).

Jeff Bezanson from the Julia project gave a short intro to Julia. What stood out to me the most was not how cool Julia looks (though it does!) but he just seemed like such a lovely guy. He said there’s lots of work to be done on Julia, and now I really want to contribute. Apparently most of the standard library for Julia is written in Julia! Also Stefan Karpinski is a Hacker School resident and <3 Hacker School. Here’s the video of his talk.

In the lightning talks, Robert Speer demoed a neat library for fixing Unicode called ftfy.

I also met Kat Chuang who runs PyLadies NYC and is great.