ISSUE 370  ·   January 18, 2022

Trends

Google Research: Themes from 2021 and Beyond

In this annual retrospective, Jeff Dean takes a look at the state of machine learning in 2021 and highlights five key trends that are poised to make big impacts around the world. This is an awesome post with lots of links to explore along the way.
Google AI Blog | Jeff Dean

Sponsored Link

How to Get Strategic Value from Data Analytics

Read this report to learn about the state of data analytics and the obstacles preventing its progress. Discover how broader accessibility and application of analytics in daily operations could help organizations produce more value. Get your free copy.

Tutorials, Projects & Opinions

Data Mishaps Night

Thursday, Feb 24th @ 7pm CST 
The second annual Data Mishaps Night features a lineup of "data mistake" stories with a focus on the human aspect of data work and lessons learned the hard way. Learn about projects that have been bungled, data roads not taken, missed commas killing pipelines, and all the rest. And if you have a mishap of your own to share, apply now! 
Data Mishaps Night

State of machine learning in Julia

Great discussion on the Julia Discourse about the current state of machine learning in Julia and where things are going. Covers things like where Julia really shines today, commonly used packages, what's lacking, issues, performance and much more.
Julia Discourse | started by Seth Axen

Using databases with Shiny

Many Shiny users know how to query existing databases but there's a lot more involved in creating and managing your own. In this step-by-step guide, Emily Riederer shows how to get started and, especially, how to think through things like security and performance.
Emily Riederer

Saving the Scranton Branch with Lead Scoring

This is a fun, introductory tutorial that shows to build a lead scoring model built using logistic regression. 
Will Crowley

PostHog: Open-source Product Analytics

Join thousands of other developers who use PostHog to build better products and analyse user behaviour. You can deploy PostHog on your infrastructure to unlock the power of open-source product analytics and session recording, essentially replacing multiple tools at once.
// sponsored

Code & Tools

xmanager

XManager is an open-source platform for packaging, running and keeping track of ML experiments. Launch experiments locally or on Google Cloud Platform (GCP) using XManager's APIs with Python.
GitHub | deepmind

Resources

Bayesian Modeling and Computation in Python

Nice introduction to Bayesian statistics aimed at helping beginner Bayesian practitioners become intermediate modelers. This is a hands-on text with examples using PyMC3, Tensorflow Probability, ArviZ, etc. 
Osvaldo Martin, Ravin Kumar, Junpeng Lao

Statistical Rethinking

This online course teaches data analysis, with a focus on scientific models first. The course prioritizes conceptual, causal models and forming precise questions about those models. The course filled up fast but slides and lectures are posted online each week.
GitHub | Richard McElreath

Data Visualization

Fooled by Beautiful Data

A "beautiful" plot isn't just about aesthetics. This preprint explores four studies that show how people commonly (mis)place trust in more beautiful graphs.
PsyArXiv | Chujun Lin, Mark Thornton

A ggplot2 Tutorial for Beautiful Plotting in R

Awesome ggplot2 tutorial with lots of examples. Includes a linked Table of Contents and useful resources at the end. This was featured in Data Elixir when it came out in 2019 but it's been substantially updated since then and it was recently plagiarized. This is the version you want!
Cédric Scherer

How did you like this issue of Data Elixir?

👎    1     2     3     4     5   👍

Was this email forwarded to you?  Sign up here >>