Runnable, tested code examples for the Agentic Data Science specialization by Pragmatic AI Labs: the same four data science problems solved in five languages with an AI coding agent in the loop, under strict quality gates.
| Course | Language |
|---|---|
| C | R |
| B | Julia |
| D | Kotlin |
| A | Rust |
| E | Lean 4 |
Specialization: https://www.coursera.org/specialization/agentic-data-science (link goes live at launch).
datasets/: Palmer Penguins and UCI Wine Quality, each with LICENSE and SHA256SUMS; fixed folds, splits and centroids infixtures/; golden reference numbers ingolden/.ontology/: vocabulary and SHACL shapes that every example'sexample.ttlmust satisfy.- Language packages (
r/,julia/,kotlin/,rust/,lean/) are added as each course's examples land.
make gate- Palmer Penguins: CC0 1.0 (Palmer Station LTER data policy).
- UCI Wine Quality: CC BY 4.0. Cortez, P. et al. (2009), doi:10.24432/C56S3T.
scripts/hero.sh draws assets/hero.svg from datasets/penguins/penguins.csv (one dot per row); rmedia svg2png renders assets/hero.png.
