ISSUE 333 · April 27, 2021

In the News

Cetacean Translation Initiative

Machine learning has enabled groundbreaking advances in human language translation so it's not surprising that research is also underway to enable inter-species communication. It's early days still but the research is fascinating. This paper introduces a road map and rationale for a new research effort that intends to decode Sperm Whale vocalizations! For more, see the related article with recordings and photos at National Geographic >>
arXiv

Tutorials, Projects & Opinions

Scaling Data: data informed to data driven to data led

Successful companies approach data as a lever for growth, not as a team to hire or a set of tools. In other words, success comes from being data led, not data informed. It's more than just how you think about it. Getting to data led requires a data maturity shift and you need to evolve key capabilities along the way. This roadmap shows how.
Reforge | Crystal Widjaja

Run Your Data Team Like A Product Team

If you're still not convinced about the limitations of being in a data-informed organization, this post takes a different approach and shows how many data teams get stuck in a service-oriented culture where the task is to answer other people's questions. Here's how that squanders the data team's real potential and what you should do instead.
Locally Optimistic | Emilie Schario and Taylor A. Murphy, PhD

A Coherent Model for the Kentucky Derby

In his latest post, Peter Cotton shows how to construct a coherent model for the upcoming Kentucky Derby. The post is based on a recent paper where he introduces a scalable approach for inferring the distributions of participant scores in a multi-party contest. Beyond the track, it's useful for things like e-commerce, auto ML, bandit-like problems, and more.
Microprediction | Peter Cotton

The missing piece of the modern data stack

Today's data stack is pretty good but is it missing a layer? Benn Stancil makes a convincing point here that a new layer for metrics would go a long way to ensure consistency for some of the most important and common use-cases for data.
Benn Stancil

We were promised Strong AI, but instead we got metadata analysis.

The late nineties dream of search engines was that they would use grand-scale AI to find everything, understand most of it and help us retrieve the best of it. Not much of that has really come true.
Cal Paterson

From spreadsheet to interactive web app in 2 minutes

Eliminate the struggle around structured data. After 6 years of leading data teams @Google, we built Polymer Search, the world's fastest, easiest and in-depth data productivity application to make you excel, regardless of skill level. Experience 100% data automation that boosts expertise and clarity for anyone in 2 mins. Get Early Access for free.
// sponsored

Resources

Introduction to Computational Thinking

This current course from MIT uses the Julia language to approach real-world problems with data analysis and computational and mathematical modeling. Lectures, interactive notebooks, and cheatsheets are all freely available. To see if it's for you, check out the short Introduction Video.
MIT

Reviews

The Art of Mathematics in Chalk

Even when it may be hard to understand, math is beautiful. Photographer Jessica Wynne set out to capture this appeal when she began photographing mathematicians' chalkboards around the world. Some of the photographs are reproduced here and have been collected in a new book: Do Not Erase: Mathematicians and Their Chalkboards.
Scientific American | Clara Moskowitz

Sign up to get Data Elixir's data science newsletter in your Inbox >>

Data Elixir is curated and maintained by Lon Riesberg. If you have questions or suggestions for the newsletter, just reply back to this email.

To find specific content from prior issues or to research topics, check out the catalogued Archives on Data Elixir's Search Page >>