| | Learn how teams move from notebooks and one-off scripts to reliable, production-ready workflows with Airflow. Whether you're building data pipelines, training ML models, or running generative AI jobs, Airflow handles the scheduling, monitoring, and orchestration so you can focus on the work itself.
*Message from this week's sponsor, Astronomer |
| | |
| | | | MicroGPT distills the algorithmic core of a GPT into a tiny model, and this interactive site lets you explore it step by step. Visualize attention, training dynamics, and architecture changes as components are added, making the transformer stack much easier to reason about.
James Tauber |
| | |
| | From Kaggle to DARPA grand challenges, this report maps the current competition landscape and digs into the techniques used by winning teams across 2025 contests. If you’re curious how top competitors approach tabular data, vision, or NLP problems, there’s a lot to learn here.
ML Contests / Research |
| | |
| | These are picks from a recent Bluesky thread. Check the thread for more! | |
| | |
| | If you’re working with spatial climate data, GAMs offer a powerful middle ground between simple interpolation and heavy physics models. This walkthrough shows how they can downscale solar radiation across the Canary Islands’ complex terrain while keeping the model interpretable.
Dr. Dominic Royé |
| | |
| | | | Most visualization books teach syntax. This one teaches judgment. Kieran Healy’s classic Data Visualization guide is getting a major update, with modern R tooling, refreshed examples, and new material reflecting years of teaching and real-world feedback. The full draft is free to read online.
Kieran Healy |
| | |
| | This site compares four popular Python visualization libraries using the Gapminder dataset. Each visualization is implemented in matplotlib, seaborn, plotnine, and altair, presented in a way that makes it easy to see how they compare across a variety of plot types. Includes code snippets for each.
Lalten |
| | |
| | | LLM coding tools are more useful when they know your ecosystem. These Claude Skills teach the model modern R practices: tidyverse conventions, Quarto workflows, Shiny dashboards, and package release checklists. Think of them as reusable expertise you can plug into your development environment.
R Works | Isabella Velásquez |
| | |
| | | In most ML workflows, humans write the code and models run the experiments. In his latest project, Andrej Karpathy flips that around. autoresearch lets an AI agent edit the training script, run experiments, evaluate outcomes, and keep iterating. Still an early prototype, but a fascinating direction.
GitHub | Andrej Karpathy |
| | |
| | | These two topic-focused editions complement the regular Tuesday issue and go deeper into the topics you’re most interested in: | Data Practice — Everyday data work, including SQL, statistics, visualization, R/Python, and tools. Published Wednesdays. Learning Machines — Machine learning and AI, including tutorials, research, tools, and applications. Published Thursdays.
| Opt-in to one or both. |
| | |
| | | This week's issue sponsored by: | | |
|
