— In the News —
This McKinsey Report provides a great overview of machine learning for smart people that aren't necessarily machine learning experts. This is really an opportunities and strategies report for the C-Suite, which provides insights into how well machine learning is understood and appreciated by decision-makers.
As part of Obama's Police Data Initiative, researchers are studying predictive analytics to identify officers whose unprofessional behavior could cause problems in the communities they serve. There are a lot of interesting issues around this.
Thought-provoking essay about freedom and control in a world driven by metadata. The internet is going to do to us whatever it is going to do - and it’s far too late to stop it.
— Sponsored Link —
— Tools and Techniques —
Nice overview of a new data analysis framework called "Ibis" that has the goal of making big data as easy to work with as small data. It’s exactly the same Python you know and love but at scale.
Comprehensive tutorial for importing data into R. This is easy to follow and covers a wide variety of file types.
Nice tutorial that explores the use of sentiment APIs to plot the shape of a story and then find other stories with similar plotlines. This is well done and includes a Jupyter notebook to make it easy to play.
— Resources —
This course is a deep dive with mathematical proofs, detailed references, and links to additional resources. Slides and videos are available in each section of the outline and a searchable/tagged collection of videos is also available at computervisiontalks.com
Curated collection of resources dedicated to recurrent neural networks. This collection includes code, lectures, books, papers, datasets, and lots of applications. If you're interested in recurrent neural networks, this is a project to watch.
— Data Viz —
Nice gallery of htmlwidgets for R. There are 52 widgets here with links to examples and documentation for things like dygraphs, threejs, d3heatmap, rbokeh, rpivotTable, sparklines, and many others. Highly recommended.
In a series of data visualizations for FiveThirtyEight, Randy Olson shows What 12 Months Of Record-Setting Temperatures Looks Like Across The U.S. Here's a link to the GitHub repo of Python code and the data that was used for the story. It's not complicated and could be useful for jumpstarting other weather-related data viz projects. After only two days, this repo has been starred over 1500 times!
— Conferences & Events —
The 2nd International Conference for Predictive APIs and Apps is coming up on August 6 & 7 at the Menzies Hotel in Sydney, Australia! Click the link above for information and definitely check out the video. It looks like a great conference!