ISSUE 423 · February 7, 2023
Profiles
How Duolingo’s AI Learns What You Need to Learn
Duolingo is known as an awesome language-learning app but the company's ambitions go much further. It's adaptive learning system, "Birdbrain," automates the traits of good tutors and it can be generalized to other topics, such as math. This article explores how Duolingo has evolved, the challenges, and where it's going.
IEEE Spectrum
Sponsored Link
You're invited!
Webinar: Modernizing Geospatial Data Analysis and Visualization with AWS & Foursquare
Business and technical leaders, learn how geospatial data can be a crucial input in understanding your market and economic landscape using AWS Data Exchange for Amazon Redshift and Foursquare Studio.
Date: February 15, 2023
Time: 10:00 AM (PST), 10:00 AM (BST), and 10:00 AM (SGT)
Tutorials, Projects & Opinions
Getting to decisions faster in A/B tests, part 1
This first post in a series explores approaches that are used in industry to get to faster decisions in A/B testing. This post is a literature review with links to lots of resources and summaries along the way.
Aurimas Račas
Examples of floating point problems
There are a lot of scenarios where floating point numbers can lead to inaccurate, inconsistent, and unexpected results. This post explores how floating point numbers work, why they aren't "bad," and walks through 8 real-world problematic examples.
Julia Evans
Should You Measure the Value of a Data Team?
Great post about the ROI of data teams. If you need to justify your team with an ROI, it's likely an organizational problem that's masquerading as a need for metrics. Metrics can help build trust with stakeholders, but they are not enough on their own. This post summarizes key arguments from several blog posts, podcasts, and discussions.
The Prefect Blog | Anna Geller
Data innovations for understanding the ocean
The Ocean Data Challenge called for data-oriented solutions to boost ocean conservation and promote a sustainable "blue economy." This post highlights the eleven winning projects, which cover things like Ocean Data as a Service, buoy-based 5G, water quality monitoring, underwater drone applications, prediction services, and more.
World Economic Forum | Ronald Tardiff
🥖 Bake data privacy and security into your product from day one by avoiding these common mistakes.
// sponsored
Tools & Code
Official MathWorks MATLAB kernel for Jupyter
This package has been around for a while but previously, it only supported the ability to access MATLAB in a browser from environments like JupyterHub. With the new update, you can now run MATLAB code in Jupyter notebooks via a MATLAB kernel for Jupyter.
The MATLAB Blog | Mike Croucher
Resources
Soccer Analytics Handbook
This popular introduction to soccer analytics by Devin Pleuler includes tutorials, links to key libraries, research papers, posts, presentations, and books. It was originally written in 2020 and has been recently updated to incorporate changes in the analytics software ecosystem. Devin Pleuler is the Director of Analytics at Toronto FC.
GitHub | Devin Pleuler
Understanding Large Language Models
Great post for getting started with large language models. There's a lot here, including links to key papers, summaries, and diagrams.
Sebastian Raschka
Data Visualization
A Better Path Toward Criticizing Data Visualizations
Data visualization critiques often take one of two approaches and neither are particularly constructive. Critiques should help expand data visualization skills and curb the spread of misleading data and graphs. This is an insightful post for achieving that.
PolicyViz | Jon Schwabish
 |