— Insight —
A lot of people seem to think that the worst of COVID-19 is behind us but that's probably far from true. After 100 days of working on the issues, DJ Patil offers insights for managing COVID-19 and ultimately, creating a data-driven public health system. Highly recommended.
A bad outcome doesn't mean it was a bad decision. In this essay, Hugo Bowne-Anderson takes a look at uncertainty, probabilistic thinking, assessing risk, and how statistics is political.
— Profiles —
Physical security at cloud data centers is more fortified than you probably think. James Bond might have a chance at getting past the first couple of security layers but getting data out of Level 5 or 6? No way.
— Tools and Techniques —
To better understand the landscape of available tools for ML production, Chip Huyen researched every AI/ML tool she could find. In this post, she explores the landscape and identifies under-served problems and opportunities. This is well-researched and insightful.
GitHub just released a collection of new tools to help with automation, collaboration and reproducibility in your data science and machine learning workflows. Here are the details.
This classic tutorial from Google's Codelabs has recently been updated for Keras and Tensorflow 2.2. It's easy to follow and is a great starting point for learning about neural networks and things like softmax, cross-entropy, learning rates, dropout and batch normalization.
— Data Viz —
In a new series of posts, Lisa Charlotte Rost explores color blindness and what it means for designing data visualizations. This is a great series with a lot of visuals that demonstrate common issues and ways to handle them.
— Career —
Here's where to find no-cost lectures, seminars and complete courses from universities like MIT, Stanford and Harvard.