ISSUE 280  ·   April 7, 2020        

Insight

All models are wrong, but some are completely wrong

Epidemiologists are making the same mistakes that the climate science community made a decade ago. Ultimately, all models are wrong. But to maintain credibility, it's important to set reasonable expectations about how wrong they might be.
Royal Statistical Society

Coronavirus Case Counts Are Meaningless*

*Unless you know something about testing. And even then, it gets complicated.
FiveThirtyEight

Sponsored Link

See what Ancestry and Autodesk use to train their models

Build better models faster. Track your datasets, code changes, models, and experimentation history . Comet provides insights and data to build better models, faster while improving productivity, collaboration and explainability. Trusted by over 10,000 data scientists globally.
Try for free today.

Tools and Techniques

An overview of ML development platforms

This guide to machine learning platforms covers a variety of options from cloud-based platforms to self-hosted studios. Entries include descriptions, screenshots, related links, and discussion of strengths & weaknesses. This is the first of a three part series.
LinkedIn | Louis Dorard

Product management for AI - what you need to know

For anyone interested in the big picture, this article by Peter Skomoroch and Mike Loukides is a must-read deep dive into the ways that AI-driven products are different than typical software products. Covers development differences, organizational prerequisites, identifying viable machine learning ideas, and how to decide which projects matter.
O'Reilly Radar

Forecasting Best Practices

Nice collection of best practice guidelines and forecasting examples from Microsoft. There's also a library of utility functions here and everything is provided as Jupyter notebooks and R markdown files.
GitHub | Microsoft

Swift: Google’s bet on differentiable programming

A team at Google has been working on making Swift the first mainstream language with first-class language-integrated differentiable programming capabilities. Here's what that means, how they chose Swift and why the project could become a big deal.
Tryolabs

🤖⚡️ Daily scikit-learn tips

As a follow-up to his popular "pandas tricks" series, Kevin Markham is now publishing a series of scikit-learn tips. You can subscribe to receive new tips as they're published or watch this repository where each new tip links its own Jupyter notebook.
GitHub | Kevin Markham

Data scientists are in demand on Vettery

Vettery is an online hiring marketplace that's changing the way people hire and get hired. Ready for a bold career move? Make a free profile, name your salary, and connect with hiring managers from top employers today.
// sponsored

Resources

Bayesian Data Analysis

The latest edition of Bayesian Data Analysis is now free to download for non-commercial purposes. Along with a download link, the landing page includes links to related course materials, demos, notes, and software. This highly acclaimed text was the winner of the 2016 De Groot Prize from the International Society for Bayesian Analysis.
GitHub | Aki Vehtari

⚽ Soccer Analytics Handbook

Great introduction to ⚽ analytics. Includes an evolving collection of Jupyter notebooks and a curated list of key research papers, posts, presentations, and books.
GitHub | Devin Pleuler

Conferences & Events

Free Metis Corporate Training Series: Intro to Python
This is your chance to give your team a strong foundation in Python by learning with Metis, a leader in data science and analytics training. Join us throughout April for a free, live online training series created especially for business professionals.  Absolute beginners welcome! No prerequisites required!  Every Tuesday in April at 12pm ET.

Data Elixir is curated and maintained by @lonriesberg. For additional finds from around the web, follow Data Elixir on LinkedIn, Twitter or Facebook.