ISSUE 261   ·   November 26, 2019        

Insight

Local-first software: you own your data, in spite of the cloud

Cloud apps are great for accessing data from any of your devices and for enabling real-time collaboration. But who owns the data? And what happens when organizations shut down and apps stop working? Here's a proposal for a new approach and there's a lot to think about. Follow this post by Adrian Colyer and start down the rabbit hole.
The Morning Paper

Global AI Survey: AI proves its worth, but few scale impact

Most companies report measurable benefits from AI where it has been deployed; however, much work remains to scale impact, manage risks, and retrain the workforce. This new report from McKinsey & Company shows how a group of high performers is leading the way.
McKinsey & Company

Sponsored Link

ML Best Practices: The Keys To Achieving Quality Data Annotation

To maximize the potential in ML models, high quality data is key. When measuring the quality of labeled data consider IOU, Accuracy, Recall, Precision, and F1 Score. An enterprise-grade labeling platform that employs a holistic annotation approach can scale quality annotation without diminishing complexity and compromising accuracy.
Learn more.

Tools and Techniques

Using Data to Analyse Team Formations ⚽

Great post by Laurie Shaw that summarizes her work on team formation analysis and shows how her new approach of measuring and classifying team formations as a function of game state is more effective than other techniques. This was recently presented at the FC Barcelona Sports Analytics Summit and won the "best research paper prize."
EightyFivePoints Blog

Cost-benefit analysis in R

This tutorial by Peter Ellis shows how to perform cost-benefit analysis in R and how to build in uncertainty in a much clearer way than is generally done.
free range statistics

How to apply machine learning and deep learning methods to audio analysis

In this tutorial, Gideon Mendels shows to analyze, explore and understand audio data using Comet’s meta machine-learning platform. This is easy to follow, starting with a introduction to audio data and digital signal processing.
Comet Blog

Metis Webinar | AI ROI: The Questions You Need to Be Asking

As business leaders increase investment in advanced analytics, data science, and AI, many struggle to assess the return on those efforts. This is due to the inaccurate measurement and reporting of success. Fortunately, we can do better. This webinar, with Q&A, will give leaders the tools to identify and assess the possible impact of data science projects.
// sponsored

Resources

Financial Models Numerical Methods

Nice collection of interactive notebooks about quantitative finance. There's also a worthwhile discussion about this project and related resources on Hacker News.
github/cantaro86

Deep Learning with PyTorch

This collection of excerpts from the upcoming book, Deep Learning with PyTorch, is a nice introduction to building and training neural networks.
pytorch org

Career

Machine Learning Systems Design

This new section of Chip Huyen's upcoming Machine Learning Interviews book offers interview strategies, linked references, case studies and exercises for thinking through machine learning system design. This is a great read for anyone who's either hiring or is interested in a career that involves production ML.
github/chiphuyen

Competitions

Hakuna Ma-data: Identify Wildlife on the Serengeti

Camera traps are an invaluable tool in conservation research, but the sheer amount of data they generate presents a huge barrier to using them effectively. In this new competition from DrivenData, you can help conservation research by building the best algorithms for wildlife detection and compete for a share of 20K USD in prizes.
DrivenData

Data Viz

Streetmaps

Easy to follow, step by step tutorial that shows how to create streetmaps for any city using osmdata and ggplot2.
ggplot2tor Blog

A Reflection on VIS2019: Or, How Doomed Are We?

Technical ability is important but if you lack grounding in fields such as psychology, stats, and communication, your visualizations won't go very far. In this post, Michael Correll summarizes key talks from VIS2019 that address non-technical challenges of communicating with data.
Multiple Views Blog

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