Scott Draves and the Beaker Notebook


By: Sheamus McGovern, ODSC Chair

Scott Draves is an extremely interesting guy to say the least. I had the privilege of interviewing him at ODSC’s Boston Data Festival.

Firstly, he’s a software artist. One glance at his work and you’ll know that’s an apt title.

Scott works at Two Sigma, an alternative investment firm (a.ka. hedge fund) with a reputation for hiring some of the smartest engineers out there. He’s the driving force behind the Beaker Notebook.

If you haven’t tried notebook style development it can be a liberating experience from your general IDE, especially if you work with data. One example: it allows you to quickly write a few lines and graph it. Speed matters!

Although it’s been around for a few years, this was my first exposure to Beaker I must say I’m impressed. As I mentioned, it’s a notebook-style development environment that’s especially good for data scientists with its neat features for working with tricky datasets, especially large ones.

Beaker is essentially language agnostic. The list it accomodates is as long as my arm and includes Python, JavaScript, Scala, Java, Python3, R, Julia, Groovy, Node, Ruby, HTML, Markdown, and LaTeX with more being added.

Beaker achieves this using auto-translation. You write snippets or “cells” in one language and that snippet is translated into another, allowing you to use all the native features of Beaker without being tied to a single language.

With half-hearted apologies to those who spend hours debating R vs Python, the folks who code in multiple languages realize there are inherent strengths and weaknesses given the domain and context.

beaker logo

Think how many times you’ve thought, as you’ve struggled with features in language X, that you could easily produce or consume/model data in language Y. Beaker allows you to leverage that language flexibility and easily glue them together in the same notebook. That’s a big plus.

This also leads to the nifty feature that allows you to run multiple independent code snippets (“cells”) written in different languages. This is very liberating from your standard IDE environment and allows for flexible, collaborative, and iterative prototyping and discovery.

Beaker also allows you to work with your chosen language’s native visualization tools but gives you the option for leveraging other data viz libraries. Tired of gplot and want to woo with D3? Beaker can accommodate. It also allows you to define your own charting API. Beaker can even open IPython notebooks and convert them to its own format.

Check out Scott’s Boston Data Festival Interview with me.  We will be publishing his talk, along with all of the other talks on our blog, so check back soon!

New to Open Data? Register for Free

    Latest Posts

    Water World


  • Related posts