Yellowbrick is an open-source Python library of visual analysis and diagnostic tools for machine learning. It extends the scikit-learn API with a suite of “Visualizers” that produce publication-quality matplotlib charts, so you can steer model selection with your eyes instead of staring at raw metric tables.
Model selection is hard to do from numbers alone. Yellowbrick’s visualizers plug into the familiar scikit-learn fit/score workflow and render what is actually happening: feature visualizers such as Rank2D compare features pairwise and rank their relationships, while model visualizers like ROCAUC plot the sensitivity–specificity tradeoff of a classifier directly. The library covers feature analysis, classification, regression, clustering and target diagnostics, all following the same consistent API of fitting a visualizer, transforming data and calling show(). It installs from PyPI or conda, includes built-in dataset loaders for examples and teaching, and is documented with an extensive gallery at scikit-yb.org. Yellowbrick began at District Data Labs, is a NumFOCUS affiliated project with a published JOSS paper, and is Apache-2.0 licensed. It has more than 4,400 stars on GitHub, with releases continuing into 2025. Because every visualizer returns a fitted matplotlib figure, the charts drop straight into notebooks, reports and papers without extra export tooling. The project also maintains teaching materials used in data-science courses.
Free. Yellowbrick is open source under the Apache-2.0 license and installs from PyPI or conda at no cost.
Data scientists and ML practitioners who build models with scikit-learn and want fast, visual diagnostics for feature selection, model comparison and error analysis, plus educators teaching machine learning concepts visually.