— In the News —
Computers have never been good at answering the type of verbal reasoning questions found in IQ tests. Now a deep learning machine unveiled in China is changing that.
Google made some big improvements to its Google Trends product this week. There's now extensive real-time data, curated datasets, and a new homepage that's all about data stories. This is definitely worth exploring and be sure to watch the video at the bottom if you're interested in where this is going.
President Obama has quietly recruited top tech talent from the likes of Google and Facebook. Their mission: to reboot how government works... Great article about how that's going.
— Tools and Techniques —
Brilliant video showing how the guts work for this machine learning algorithm. It's just 6 minutes and if you want more, there are links to the source code and a paper. Highly Recommended.
This post in Google's research blog explores some simple techniques for looking inside neural networks. The results are fascinating and have generated a lot of discussion around the web this week. The article is easy to follow and it's also worth checking out the discussion in Hacker News where it had over 700 upvotes within a day of being posted.
Packages are the fundamental units of reproducible R code. They include reusable R functions, the documentation that describes how to use them, and sample data. If you use R, definitely check out this site by Hadley Wickham that goes along with his new book R Packages. The site includes the complete text and downloadable code.
— Resources —
Nice collection of curated iPython notebooks for Data Science. These notebooks cover topics for Spark, Hadoop MapReduce, HDFS, AWS, Kaggle, scikit-learn, matplotlib, pandas, NumPy, SciPy, and various command lines.
Online text for a computational statistics course at Duke University. The text is Python-based but there are lots of code samples and it should be easy to follow even if you're not a Python pro. This is definitely worth bookmarking.
— Data Viz —
This article explores seven ways to tell the story of a single dataset. It's not exhaustive but there are good ideas here for creating your own data narratives.
— Archive Pick —
And to make sure that your data stories are presented accurately, make sure you're aware of the visual tricks in this article. Some of these would be easy to stumble into and could result in unintended spin.
— 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: