— In the News —
Startups are keeping head counts low, and even eliminating management positions, by replacing decision-makers with data. Here's how that's going.
Bloomberg Philanthropies announced this week that it is investing $42 million to create the What Works Cities Initiative, which is aimed at helping 100 mid-size cities make better use of data in their policymaking. What they're hoping to accomplish might surprise you.
In 2013, the U.S. Defense Advanced Research Projects Agency (DARPA) launched a four-year program to fund research in probabilistic-programming. These techniques are designed to make machine-learning applications easier to build. While "50 lines of code" is an unfortunate distraction, this is a good overview of the research and its progress.
Great article in Wired about how computer vision is the key enabling technology for all of AI.
— Tools and Techniques —
Fantastic collection of 55 useful tools for data visualization. This collection is curated by the folks at Datavisualization.ch and is well-organized and easy to search. Highly Recommended.
StockTwits is a Twitter-like service that allows users to post news, ideas, and questions about individual stocks. It also supports sentiment tags, to indicate whether a user is bullish or bearish about a given stock. Building a stock trading algorithm based on those tags can be surprisingly, quite lucrative. This article explores the approach and offers sample code and data to play with.
Nice overview of markup languages that make it fast and easy to include documentation with your projects. Whether you want to create HTML, ePub, Mobi, PDF, Python documentation, or even interactive code, you'll find an option here to streamline your workflow. This is well-written and includes lots of examples and links to references.
— Resources —
This 8-page PDF includes a comprehensive suite of notes summarizing important probability concepts, formulas, and distributions, with examples, stories, and solved problems.
Robin Wilson from the University of Southampton maintains this list of freely available geographic datasets. There are over 300 sites listed here that cover an amazing breadth of categories including things like elevation, weather, climate, natural disasters, ecology, wars, population, points of interest, roads, and government data from throughout the world. This is definitely worth bookmarking.
This collection of data science and machine learning podcasts generated a lot of interest around the web this week. This list includes a couple of podcasts I hadn't heard of and is a solid list with good descriptions and links to each.
— Inspiration —
This might be a stretch to include in Data Elixir but what a fascinating stretch it is. This behind-the-scenes look at the new movie, Ex Machina, is a MUST SEE.