| | Transform data into insights with the online Master of Translational Data Analytics from Ohio State’s Translational Data Analytics Institute. This interdisciplinary program blends machine learning, user experience, data visualization, and design thinking to prepare you to uncover and present insights from data.
*Message from this week's sponsor, The Ohio State University |
|
| | | This slide deck is one of the most comprehensive overviews you'll find about the current state of AI. It highlights developments in research, industry, policy, and safety, plus predictions for the coming year. It’s dense, but the slide format makes it easy to browse.
Nathan Benaich and Air Street Capital |
| | |
| | 🔗 Note: if you don’t see links in this email, it’s also available on the web: dataelixir.com/archive/issue-593.html |
| | |
| | Statistical rules of thumb are easy to remember but for them to be useful, you need to know when to trust them. In this post, Jim Frost walks through seven familiar shortcuts, including the rule of 30 and the 80/20 principle, explaining where they come from, when they help, and how much confidence they deserve.
Jim Frost |
| | |
| | A DuckDB database doesn't need to hold any data to be useful. This walkthrough shows how to store a database of views that point to remote Parquet files, which essentially turns it into a tiny, shared catalog. Users get ready-to-query datasets through a single URL, while the data stays where it is.
The DuckDB team |
| | |
| | Sampling can make a query much faster, but it matters how you pick the rows. In this post, Radim Marek explains PostgreSQL’s sampling options, why nearby rows can give you less information than you expect, and how to build consistent samples that still work across related tables.
Radim Marek |
| | |
| | AI can now outperform licensed accountants on certain detailed accounting tasks, but that doesn’t mean it can do the whole job. Using tasks adapted from APEX-Accounting, Mercor explores what these results suggest about the future of work and the growing importance of communication, context, and asking the right questions.
// sponsored |
| | |
| | | | Edward Tufte’s data-ink ratio encourages people to remove chart elements that don’t convey information, but how far should you go? This interactive guide lets you adjust bars, labels, and other details to see what makes a chart more clear and what gets lost along the way.
Science UX | Tufte's Razor |
| | |
| | Sometimes a few points on a chart need more explanation than a short label can provide. ggindex is a new ggplot2 extension that marks those points with numbers and puts their descriptions in a matching legend. That gives you room to explain them without crowding the chart.
Discindo |
| | |
| | | Polars 2.0 brings some useful changes for everyday data work, including streaming by default, the ability to spill some operations to disk when memory runs short, and expanded SQL support. If you’re upgrading, check the migration guide, especially if your analysis depends on row order.
Polars | Ritchie Vink |
| | |
| | R has several new tools for managing packages and environments and it can be hard to keep them straight. In this post, Isabella Velásquez explains how five options differ, with examples for sharing a project, recreating its software environment, or making a single script easier to rerun.
R Works | Isabella Velásquez |
| | |
| | | | This new site compresses climate charts into visual summaries for cities around the world. Click through for city-specific pages that show 30-year climate normals, covering things like temperature, precipitation, humidity, cloud cover, wind, and more. The code is open-source and it's a nice example of clean visual design.
Weather Ribbon |
| | |
| | | This week's issue sponsored by: | | |
|
