Definition
Plain language
A free Python toolkit for working with big labeled tables of scientific data.
As stated in the literature
A Python library for labeled multi-dimensional array data, widely used in climate and geoscience work; appears as a source of real-world bugs in software-engineering agent benchmarks and dynamic-analysis debugging case studies.
Why it matters: It matters because it's a widely relied-on tool for scientific data, and its real bugs make it a useful proving ground for debugging agents.
For example, a climate scientist can use xarray to slice a temperature dataset by year and region using clear labels instead of raw index numbers.
Heard on the show
“There's a case study with xarray — that's a scientific data library in Python.”Episode 142 — Training a Tiny Model to Run the Plumbing Between an Agent and the World