Issue 55
— In the News —
AI Is About To Go Mainstream And Reshape The Workplace
AI-powered apps will be a gold rush for entrepreneurs. Here's a survey of the landscape.
Learning to learn, or the advent of augmented data scientists
How much of what data scientists do could be automated? This is a nice overview of the problems and opportunities.
Quantifying my Transition from Academia to Data Science
Great exploration of a postdoc's path to a non-academic job. By meticulously tracking how she spends her time, Alex Smolyanskaya provides valuable insights into the differences between academia and industry.
Fantastic Reddit AMA with Hadley Wickham. There are lots of useful insights, links, and discussions here. Highly recommended. Which is better: a baseball player who's achieved 4 hits in 10 chances or one with 300 hits in 1000 chances? This post isn’t really about baseball. It's about a very useful statistical method for estimating a large number of proportions, called empirical Bayes estimation. This is a fantastic explanation. Here's a gentle introduction to neural nets with an interactive visualization to help show how they learn. Great Data Journalism tutorial by Lena Groeger. This flows very well and includes many useful references and examples.
— Tools and Techniques —
I'm Hadley Wickham, Chief Scientist at RStudio and creator of lots of R packages. Ask me anything!
Understanding empirical Bayes estimation (using baseball statistics)
How do neural networks learn?
Intro to Data & Code
Nice list of 150+ blogs, organized across eight different categories: machine learning, Hadoop, R, Python, business intelligence, data visualization, statistics, and a general data science bucket. This set of Notebooks is written for scientists and engineers who want to use Python for exploratory computing, scripting, data analysis, and visualization. Each notebook covers a specific topic and includes exercises, data, and an accompanying video.
— Resources —
The Ultimate Guide to Data Science Blogs
Exploratory Computing with Python
No spam, ever. For a sample issue, check out the latest: #280