Altair Ally

This package is in early development and its API might change without notice.

Altair Ally is a companion package to Altair, which provides a few shortcuts to create common plots for exploratory data analysis (EDA), particularly those involving visualizing an entire dataset. The goal is to encourage good EDA habits by making it quick and simple to create these visualizations, and to provide useful default interactions with the plots.

Thanks to the excellent API in Altair / Vega-Lite, some configuration is possible by using the built-in methods of the returned chart objects, but the general philosophy is that highly customized plots are better made in Altair directly. Check out the examples section to get started and the API reference for all available options.

The package name is a nod to ggally which complements ggplot2 in a similar, but much more extensive manner.

Installation

pip install altair-ally

Version 0.2.0 requires Python 3.11 or newer and supports Altair 6.x, including 6.3.0, and pandas 1.5.3 through 3.x. Examples use Altair's built-in dataset loader, altair.datasets.

To work with a source checkout, install uv and run:

uv sync --locked
uv run python

Documentation

This Jupyter Book contains the documentation with more info and examples.

See the compatibility and migration guide for the expanded dist() options and changes since 0.1.1. See CHANGELOG.md for the 0.2.0 release notes.

Development

uv sync --locked --all-groups
uv run task test
uv run --group doc task doc-build
uv run task build

The test suite validates chart specifications and renders them with Vega-Lite. CI covers the minimum dependencies, Altair 6.0 and 6.3, and pandas 3.

Project metadata, development dependencies, pytest configuration, and tasks live in pyproject.toml. The committed uv.lock makes development reproducible; .python-version selects Python 3.13 locally, while CI also checks Python 3.11. See CONTRIBUTING.md for dependency updates and CI overrides.