— In the News —
Fantastic story about the Madden NFL video-game franchise. There's a bit of everything you'd want in a data science drama: algorithms, politics, data visualizations, clips from the game, and interviews with the stars. Highly recommended.
Since age 15 or so, Jürgen Schmidhuber's main scientific ambition has been to build an optimal scientist through self-improving Artificial Intelligence, then retire. He's still working on that but in the meantime has accomplished an enormous amount. This is a great Reddit AMA (Ask Me Anything).
The ninth annual MIT Sloan Sports Analytics Conference (SSAC) was the biggest meeting yet of sports-data experts. One thing that's clear in this synopsis is that there's a lot of untapped opportunity in the sports analytics field.
— Tools and Techniques —
Awesome introduction to digital signal processing. It's based on a collection of interactive visualizations, which make it super easy to understand. Follow the index at the top of the pages or look for the "Next" button at the bottom to navigate. Even if you don't think you're interested, this is a MUST-PLAY-WITH tutorial.
Nice tutorial for getting started with machine learning using the Python package, scikit-learn.
A month ago, Randy Olson blogged about using a variety of machine learning tricks to find Waldo. Inspired by Randy's work, Vitaliy Kaurov from Wolfram dove into the problem and examined some additional approaches. This is a great exploration with lots of diagrams and code snippets.
Why Julia? "It's concise, easy to understand, easy to extend, and reasonably efficient." Great tutorial.
— Resources —
You'd be hard-pressed to find these datasets anywhere else. This curated collection includes things like twitter sentiment on a variety of topics, labeled images, and crowdsourced colors in 9 languages. All datasets have been cleaned and are available as csv files. Definitely worth bookmarking.
— Data Viz —
This chart made the rounds on the Internet recently and drew a fair bit of criticism. This is a worthwhile critique and offers some alternative ways of presenting the same data.
— About —
Data Elixir is curated and maintained by @lonriesberg. If you find this newsletter worthwhile, please help spread the word! Forward to your colleagues or Click here to Tweet. Thanks.