— Insight —
On its own, data isn't valuable. In this deep dive, James Currier shows how data network effects can be powerful and why they're often misunderstood.
The 2020 Magic Quadrant Report explores the enterprise software landscape for data science and machine learning applications. This is a great overview of the key players and how the market is evolving.
— Tools and Techniques —
Python's built-in itertools and more-itertools provide "the whole kitchen sink" for processing and iterating over data in Python but they're not widely known. In this post, Martin Heinz explores some of the most useful functions in these libraries, including lots of examples.
This curated learning path by Neil Sainsbury covers a wide range of topics, including foundations, mathematical reasoning, linear algebra, probability, and lots more. You'll mostly find books here but there are also a handful of online courses, YouTube channels, and communities.
This is an awesome introduction to Computer Vision using just Excel. After walking through basic techniques, the tutorial shows how to implement and visualize more advanced algorithms such as Face Detection and Hough Transform without using external scripts or plugins.
Great post for learning how GPT-2 works. In this post, Aman Arora walks-through the Huggingface implementation of GPT-2 and explains the major components along the way. Follow the links too!
Fastpages offers everything you'd expect from a modern blogging platform, including search, comments, markdown and extra features for incorporating Jupyter Notebooks. It's free, there's nothing to install, and setup is automated.
This guide by Naren Thiagarajan, co-founder of FloydHub, offers practical advice for optimizing machine learning infrastructure that uses AWS EC2. There are a lot of good ideas here that will be particularly useful for decision-makers and infrastructure leads.
Early Price ends this Friday 28 Feb for the O’Reilly Strata Data & AI Conference in London. Come see the latest trends in data and AI. We'll be bringing together innovators to cover data science, machine learning, artificial intelligence, and more in London from 20-23 April.
— Data Viz —
The Data Wrapper Book Club is starting up again and this time, instead of a book, the discussion will be focused on 6 data visualization papers. This post introduces the format for the discussion, the overall theme, and the papers with download links.
This package really struck a chord with people on Twitter recently. patchwork makes it super simple to combine separate ggplots into the same graphic. In that sense, it's similar to other packages but patchwork uses an API that encourages exploration and scales to complex layouts.