A data scientist's blog, but make it executable

What it does

Randal Olson’s monorepo collects teaching materials, code, and data for his data analysis and machine-learning projects. Each directory typically corresponds to a blog post on his website and contains a narrative Jupyter notebook plus supporting files. The instructional text is licensed under CC BY 4.0 and the code under MIT.

The interesting bit

The repository treats the notebook as the primary publication format. Analyses are packaged as self-contained stories that can be browsed on nbviewer or executed locally, effectively turning a blog archive into a hands-on curriculum.

Key highlights

• 6,705 stars suggest the notebook-as-blog-post format has found an audience
• Topics span evolutionary algorithms and applied projects like wheres-waldo-path-optimization
• Dual-licensed: Creative Commons for prose, MIT for software
• Viewable via nbviewer without a local Jupyter installation
• Python 2.7 and 3.5 badges indicate broad (if dated) runtime support

Caveats

• The README is strictly an index; you must enter individual directories to find specific documentation, data usage terms, and notes
• Python 2.7 and 3.5 support badges suggest some notebooks target older dependency stacks

Verdict

Worth bookmarking if you learn by dissecting real-world notebooks. Skip it if you need a unified framework or actively maintained library; this is a curated scrapbook, not a product.