— In the News —
Fascinating playlist of TED AI talks for smart, non-experts. There are 7 talks to choose from that range from 10-20 minutes each. Highly recommended.
An automated email responder from Google Research uses recurrent neural networks to decide what to write. It's early still but apparently, the initial results are pretty good. Here's a high-level view of how it works and when you can expect to use it. It's sooner than you think.
Magical things can happen when city governments make their data easily available. In Louisville, KY, the American Printing House for the Blind used the city's “open by default” policy to create an innovative audio map of the city. What they learned offers useful lessons for cities everywhere.
— Tools and Techniques —
Nice collection of lesser known ways of using Jupyter notebooks.
If you've been curious about Deep Learning but didn't know where to start, read this post on NVIDIA's developer blog! This is a gentle introduction to Deep Learning that shies away from math and theory and covers the most important concepts using analogies and images.
— Resources —
Here's a fantastic guide for getting started with machine learning. It's somewhat of a tutorial but mostly, it's a well-organized tour of worthwhile resources around the web
Large collection of Python wrappers for accessing a variety of services and data APIs. If you know what you're looking for, you could just use a search engine to find specific APIs but having them together like this is super useful for generating project ideas.
— Data Viz —
Nice synopsis of last week's IEEE VIS conference by Jen Christiansen at Scientific American. IEEE VIS is enormous in both its breadth and importance. Jen does a great job here by not trying to cover everything and focusing instead on some key gems.
There have been some beautiful visualizations of chess data being shared around the web recently. This is a great post by Joshua Kunst that describes some of those visualizations, including the code to make them.
Inspirational talk by Georgia Lupi about connecting with data on a personal level. Ultimately, the whole point of Big Data is to make things smarter, smaller, and understandable - to people. This talk offers an important perspective that's often under-appreciated by those of us caught up with the latest technologies.
— 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 share on your favorite network: