Some Machine Learning and Data Science Resources

I hope that, if nothing else, after taking this course you will have a continued interest in machine learning and data science. Here are a quick list of resources that are a combination of references and content aggregators that will help keep you “up-to-date.” Note that these are very opinionated, in other words, these are things David likes. These are by no means exhaustive. Just some things to think about. Maybe I should post this somewhere and keep it updated…

Books

• A First Course in Probability
  • A more detailed introduction to probability than that seen in STAT 400.
• Introduction to Probability
  • David’s personal favorite. Nicely covers partitions and conditioning.
• Neural Networks and Deep Learning
  • Often cited. Develops everything from the ground up, with code.
• Deep Learning
  • “The” deep learning book. Rather technical, not super practical.
• R for Data Science
  • A more modern approach to R.

Courses

• Stanford, CS 231n: Convolutional Neural Networks for Visual Recognition
  • The go-to course for quickly getting up to speed with deep learning after understanding machine learning. (May be helpful to first understand some optimization.) Focuses on computer vision, but is actually very general.
• Coursera: Machine Learning
  • “The” machine learning course. Much more focus on optimization than our course. Great, but a big downside is the use of MATLAB. (Unless they’ve updated to port things to Python. I hope they do that.)
• MIT: Linear Algebra
  • The instructor, Gilbert Strang, is simply wonderful. A great refresher of linear algebra.
• MIT: Probabilistic Systems Analysis and Applied Probability
  • Course to go with the MIT probability book above.
• fast.ai: Practical Deep Learning For Coders
  • A “practical” and quick introduction to deep learning.

Aggregators

• Hacker News
  • A bit of an odd culture, but a good source of articles.
• Reddit: Machine Learning Subreddit
  • Decent despite being part of Reddit.
• Reddit: Data is Beautiful Subreddit
  • Decent despite being part of Reddit.

Twitter People

If you follow a few people on Twitter, you end up staying pretty up to date with what’s going on. Remember these are people though, so they’re tweeting opinions. What they write is not gospel, but it is useful to know what is being discussed in the community.

• Hadley Wickham
  • All-around R wizard. You’ve certainly used his packages. Works for RStudio.
• Jenny Bryan
  • An early and leading data science educator. Now associated with RStudio
• David Robinson
  • Data scientist at DataCamp, formerly StackOverflow. Also maintains a good blog. Famous for analyzing Trump’s tweets.
• Hilary Parker
  • Data scientist at StitchFix, formerly Etsy. Has a podcast. Very much interested in data analysis.
• Andrej Karpathy
  • Former instructor of CS 231n. Now Telsa, formerly OpenAI.
• Ian Goodfellow
  • Leading academic in deep learning. Literally wrote the book. Google Brain.
• Szilard Pafka
  • Shares insightful opnion on practical data science and machine learning.

Youtube Series

• 3Blue1Brown: Essence of Calculus
• 3Blue1Brown: Essence of Linear Algebra
• 3Blue1Brown: Neural Networks

Podcasts

• Not So Standard Deviations
  • Podcast from an industry data scientist, Hilary Parker (StitchFix) and an academic data scientist, Roger Peng. (JHU, Coursera Data Science)

Blogs

• Variance Explained
  • Previous mentioned blog from David Robinson.