Yellowbrick

1.5 · active · verified Mon Apr 13

Yellowbrick is a Python library that extends the scikit-learn API with a suite of visual analysis and diagnostic tools for machine learning. It allows users to visualize model performance, feature relationships, and evaluate model selection processes directly within their existing scikit-learn workflows. The library is currently at version 1.5 and historically has had a release cadence of several major updates per year, though the latest release (v1.5) is from August 2022.

Warnings

breaking The `poof()` method for rendering visualizers was deprecated in v1.0.1 and subsequently removed. It was replaced by `show()`.
Fix: Replace all calls to `visualizer.poof()` with `visualizer.show()`.
breaking Yellowbrick dropped support for Python 2.x with the release of v1.0.
Fix: Ensure your project uses Python 3.6 or higher (v1.5 requires >=3.4, <4).
gotcha Yellowbrick frequently updates to maintain compatibility with rapidly evolving upstream libraries like scikit-learn, NumPy, SciPy, and Matplotlib. Specific versions of these dependencies can be required.
Fix: Always check Yellowbrick's `requirements.txt` or `setup.py` on its GitHub for exact dependency versions if encountering compatibility issues. Using `pip install --upgrade yellowbrick` often helps fetch compatible versions automatically.
breaking The internal `set_params` and `get_params` API for `ModelVisualizers` changed to align with `scikit-learn` v1.0+.
Fix: If you have developed custom visualizers or integrations that directly interact with `ModelVisualizer`'s parameter handling, review the base class implementation in `yellowbrick.model_selection` to adapt to the new pattern.
gotcha The `nltk` library and its associated data are required for text-based visualizers (e.g., `FreqDistVisualizer`, `WordCorrelationPlot`) but are not installed by default with `pip install yellowbrick`.
Fix: Install with `pip install yellowbrick[text]` or separately via `pip install nltk`. You may also need to download `nltk` data, e.g., `import nltk; nltk.download('stopwords')`.

Install

pip install yellowbrick Install base package
pip install yellowbrick[text] Install with text processing dependencies (e.g., NLTK)

Imports

KElbowVisualizer

from yellowbrick.cluster import KElbowVisualizer

ROCAUC

from yellowbrick.classifier import ROCAUC

ResidualsPlot

from yellowbrick.regressor import ResidualsPlot

FeatureImportances

from yellowbrick.model_selection import FeatureImportances

show
```
visualizer.show()
```
The `poof()` method was deprecated in v1.0.1 and removed in later versions; use `show()` instead.

Quickstart

This example demonstrates how to use the `KElbowVisualizer` to determine the optimal number of clusters for a KMeans model using synthetic data. It fits the visualizer to the data and then displays the resulting elbow plot.

from yellowbrick.cluster import KElbowVisualizer
from sklearn.cluster import KMeans
from sklearn.datasets import make_blobs

# Generate synthetic dataset
X, y = make_blobs(n_samples=1000, n_features=12, centers=8, random_state=42)

# Instantiate the clustering model and visualizer
model = KMeans(random_state=42, n_init=10) # n_init added for KMeans > sklearn 1.2
visualizer = KElbowVisualizer(model, k=(2,12))

visualizer.fit(X)        # Fit the data to the visualizer
visualizer.show()        # Finalize and render the figure

view raw JSON →