Data Science in Ruby? Is it possible? Is it Fast? Should we use it?
summary
Ruby can do data science but performance and ecosystem gaps matter: daru and nmatrix (SciRuby) provide data frames and in-memory matrices but nmatrix is slow for large data and a four‑year bug remains open; numo/narray is much faster but lacks daru compatibility and sparse matrices; red‑data‑tools uses Apache Arrow for fast I/O but lacks manipulation features. Benchmarks show numpy often ~57–100× faster than Ruby arrays. PyCall lets Ruby call numpy/pandas with native speed, and Kiba handles ETL; ruby‑spark is effectively dead. Decision tree example works but compares strings, making it brittle.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:57What Data Science Means
- 2:06Can Ruby Do Data Science
- 8:22Ruby Spark
- 19:40Wrapper Libraries for Pandas and Numpy
- 24:19The Best Work Pattern for Data Science in Ruby
