ISSUE 406  ·   September 27, 2022

In the News

‘Very Harmful’ Lack of Data Blunts U.S. Response to Outbreaks

In a perfect world, healthcare data could be easily used to optimize treatment and track disease. But in the U.S., problems abound: key data is stripped from records to help anonymize it, disparate systems don't talk with each other, tech stacks are outdated. And on and on.
New York Times | Sharon LaFraniere

How squishy math is revealing doughnuts in the brain

Analyzing squishy data? Use squishy math! This article takes a look at topological data analysis, which offers a radically different perspective than typical data analysis techniques. Topology's usefulness is its ability to reveal qualitative, rather than quantitative, properties and it identifies the aspects of a structure that persist even if there are random errors, or noise, in the underlying measurements. Welcome to squishy math.
Scientific American | Kelsey Houston-Edwards

Sponsored Link

Watch The Iron Analyst [2022] - Powered by Bright Data

Seven data scientists analyze & visualize a full data set and present their dashboards to a panel of judges. Watch now to find out who will be the next Iron Analyst!

Tutorials, Projects & Opinions

Design Principles for Data Analysis

Design thinking is a problem-solving process that helps to understand the people that a solution is being designed for. This paper introduces the idea of design thinking as applied to data analysis and offers a set of principles to help guide the development of data analyses, including their construction, methods, tooling, and workflow.
Lucy D’Agostino McGowan, Roger D. Peng & Stephanie C. Hicks

ML for Fraud Detection in Streaming Services

This latest post on the Netflix Tech Blog is a good primer on using machine learning for anomaly detection at scale.
Netflix Technology Blog

Introduction to Python and NumPy 

This notebook tutorial covers foundational topics for Python (e.g. intro to Python operations, data visualization, debugging) and NumPy data structures (e.g. scalar, vector, matrix, tensor).
DeepMind

Code & Tools

R-universe Powersearch

The next time you're looking for a specific R function or package, check out the new drop-down menu of Search options on the R-universe search page. Search by keywords, dependencies, exports, author, etc.
rOpenSci

Modin

Modin is a drop-in replacement for pandas that lets you speed up your workflows by ensuring that pandas uses all of your cores. Modin works especially well on larger datasets, where pandas becomes painfully slow or runs out of memory.
GitHub | Modin-Project

Parsr - Turn your documents into data

Parsr is a minimal-footprint document cleaning, parsing and extraction toolchain which generates enriched structured data. Works with a variety of document types, including PDF, docx, email and images.
GitHub | AXA

Career

The Hierarchy Is Bullshit (and Bad For Business)

Great rant on career paths and organizational hierarchy. Here's why "higher up" isn't always better, what you should do instead, and ideas for making your organization healthier.
Charity Majors

Discussions

This is a rabbit hole of gems for spreadsheets lovers.

How do you explore large datasets? There's a huge diversity of responses here, including lots of tools and links along the way.

How did you like this issue of Data Elixir?

👎    1     2     3     4     5   👍

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