— In the News —
This is more than a history of data science at the New York Times. This talk by Chris Wiggins dives into the roots of "data science" as a discipline and where things are going. Great talk.
Neural nets have been remarkably successful at things like drug discovery and image recognition but they can be extremely complex and difficult to understand. This article explores the issues and implications of intelligence that even experts don't completely understand.
The technology team at The New York Times recently redesigned its article recommendation engine. This article explains why and examines the algorithm decisions that were made along the way.
— Tools and Techniques —
Kalman filters can be used in any place where you have uncertain information about some dynamic system, and you can make an educated guess about what the system is going to do next. This is a great run-down of how they work and how to use them. For more detail, check out this Kalman Filter textbook, published as a set of IPython Notebooks.
The PyData ecosystem is growing rapidly with existing tools maturing and new tools appearing on a regular basis. In this talk from PyData, Rob Story examines this crowded ecosystem and brings some clarity when deciding on which tool to use in a given situation. Also, check out the IPython notebook that goes along with this talk.
This tutorial describes 10 common machine learning algorithms and includes code snippets for each in both R and Python.
Super cool experiment and tutorial. The recordings are amazing! This is very well explained and there’s also a GitHub repo of code to get started with.
— Resources —
Lots of worthwhile blogs here. The list is available via GitHub, RSS, or you can download an opml file.
— Data Viz —
There’s a lot more to using color effectively than most people realize. This article by the Plotly team breaks down some of the most important considerations for making sure that your visualizations clearly communicate your data.
Nice tool for generating palettes of optimally distinct colors.
— 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 use the links below to share to your favorite network: