abax

License: GPL-3.0 · source

abax is a statistics and data-science workstation built on a fast, scriptable spreadsheet. Load a dataset and explore it with 642 formula functions — statistics and probability distributions, financial, engineering, database, and a set built for radio work. Run the built-in analyses: regression, t-tests, analysis of variance, correlation. Reshape with pivot and group-by, profile columns, chart the result, hand a selection to pandas, and script the whole thing with Python macros.

It grows out of real practice — data science in Python and R as they are used in epidemiology, public health, and biomedical engineering research. That kind of work outgrows a spreadsheet quickly but doesn't want to become a notebook server; abax aims at the middle. Formulas behave the way spreadsheet hands expect — array results spill across cells, errors are values rather than crashes, LET and LAMBDA work — and where behavior is ambiguous, abax follows gnumeric.

It reads and writes by extension: CSV, Excel, Parquet, SQLite, JSON, Markdown, R, OpenDocument, and Jupyter notebooks with a lossless round-trip — and abax itself runs as a Jupyter kernel when a notebook is where the work lives. It runs as a desktop application, a vim-style terminal interface friendly to an SSH session, or a headless command line. The core is pure standard-library Python; the heavier machinery — hypothesis tests, machine-learning models, linear algebra, Fourier analysis, differential-equation solvers — arrives as optional pieces with graceful fallbacks, so nothing breaks when a library is missing.

The assistive-technology background shows here too: interface elements carry screen-reader labels, cells can be read aloud as you move through a sheet, and a dyslexia-friendly font is one setting away.

And because of who built it: more than sixty radio-frequency functions, link budgets, a Smith chart, and thin-wire antenna modelling live in the same spreadsheet as the statistics.

Python — pip install abax, or ready-to-run downloads on every release for machines without Python. Code at github.com/leavesofgrass/abax; full documentation at leavesofgrass.github.io/abax.