Open Textbook • Statistics & Data Science

Data Science at Nushell

A rigorous, Nushell-first open curriculum for end-to-end data science in the Linux terminal. Featuring polyglot pipeline integration with Python, R, and specialized CLI machine learning engines inside a reproducible Docker environment.

Download English PDF (5.5 MB) Download Spanish PDF (5.6 MB)

English Edition

The English release of the curriculum, adapted with standard international computational statistics terminology. Ideal for English-speaking courses and global practitioners.

Compiled PDF • Single-part 10-chapter monograph

View / Download PDF

Spanish Edition

The original Spanish edition (Ciencia de Datos en Nushell). Complete 10-chapter curriculum with code examples, pipelines, and exercises.

PDF oficial en Español • Monografía integral de 10 capítulos

Ver / Descargar PDF (Español)

Quickstart with Docker (`dsin`)

Every pipeline, script, and code block in the curriculum is 100% verified to run reproducibly inside the standalone kurokirasama/dsin:latest Docker container with pre-configured Nushell plugins, dataset mount points, and modeling engines.

docker run --rm -it -v "$(pwd)/Files:/data" kurokirasama/dsin:latest

Curriculum Structure

Chapter 1

Introduction

The terminal as a scientific laboratory, the Nushell structured data paradigm, and reproducible environments.
Chapter 2

Software Installation

The standalone dsin Docker container, interactive setup, and environment verification.
Chapter 3

Nushell Fundamentals and Data Structures

Terminal introduction, shells vs. Nushell, strings, lists, records, tables, cell paths, ttt editor, and Git.
Chapter 4

Programming and Creating Command-Line Tools

Custom commands (def, export def), flags, strong typing, modules, Google Maps API, and R/Python integration.
Chapter 5

Data Acquisition

HTTP requests (http get), decompression, Excel spreadsheets (.xlsx), SQLite databases (query db), and REST APIs.
Chapter 6

Data Cleaning and Transformation

Native Nushell filtering, structural mutations, tabular reshaping, and the high-performance Polars plugin.
Chapter 7

Exploratory Data Analysis and Visualization

Descriptive statistics (math, polars summary), polyglot scripts (.R/.py), and in-terminal plotting with timg.
Chapter 8

Serial and Parallel Data Pipelines

Serial batching (each), native concurrency (par-each), and GNU parallel orchestration.
Chapter 9

Polyglot Data Modeling

Dimensionality reduction (PCA, t-SNE, UMAP), clustering (K-Means/Hierarchical/DBSCAN/PAM), regression (VW, LightGBM, DuckDB OLS), and classification including neural networks (ml-benchmark, skll, Rscript, python3 → Nushell tables).
Chapter 10

Conclusion: The Command Line as a Data Intelligence Laboratory

OSEMN synthesis in typed shells, software engineering rigor in statistical workflows, out-of-core and zero-copy Apache Arrow architectures, and production field guidelines.

Citation

If you cite or reference this work in research, teaching, or technical documentation, please use the corresponding citation:

English Edition:
Gómez Guzmán, L. (2026). Data Science at Nushell. https://kurokirasama.github.io/data_science_at_nushell/

Spanish Edition:
Gómez Guzmán, L. (2026). Ciencia de datos en la terminal Linux con Nushell. https://kurokirasama.github.io/data_science_at_nushell/es/