Implementation:DistrictDataLabs Yellowbrick PCA Visualizer

Knowledge Sources	Yellowbrick Yellowbrick Docs
Domains	Machine_Learning, Feature_Analysis, Visualization
Last Updated	2026-02-08 00:00 GMT

Overview

Concrete tool for PCA-based dimensionality reduction and scatter plot visualization provided by the Yellowbrick library.

Description

The PCA visualizer (aliased as PCADecomposition) produces a two- or three-dimensional principal component plot of data projected onto its largest sequential principal components. Internally, it constructs a scikit-learn Pipeline containing a StandardScaler (controlled by the scale parameter) followed by a sklearn.decomposition.PCA transformer. The resulting scatter plot can be colored by discrete class labels or continuous target values. Optional features include biplot arrows (showing feature contributions to each component) and a heatmap below the scatter plot displaying the magnitude of each feature in the principal components.

Usage

Use the PCA visualizer when you want to reduce a high-dimensional feature space to 2D or 3D for visual exploration. It is compatible with both classification (discrete target) and regression (continuous target) problems. Enable proj_features=True for biplot arrows or heatmap=True for the component contribution heatmap. Note that the heatmap is not compatible with 3D projections.

Code Reference

Source Location

• Repository: yellowbrick
• File: yellowbrick/features/pca.py
• Lines: PCA class at L43-458, quick method at L466-627

Signature

class PCA(ProjectionVisualizer):
    def __init__(
        self,
        ax=None,
        features=None,
        classes=None,
        scale=True,
        projection=2,
        proj_features=False,
        colors=None,
        colormap=None,
        alpha=0.75,
        random_state=None,
        colorbar=True,
        heatmap=False,
        **kwargs
    ):

Import

from yellowbrick.features import PCA
# or equivalently:
from yellowbrick.features.pca import PCADecomposition

I/O Contract

Inputs

Name	Type	Required	Description
ax	matplotlib Axes	No	The axes to plot on. If None, current axes are used or generated.
features	list	No	Feature names. Inferred from DataFrame columns if not provided.
classes	list	No	Class labels for the legend (discrete target only). Inferred from y if not provided.
scale	bool	No	Whether to scale data with StandardScaler before PCA. Default True.
projection	int	No	Number of dimensions to project into: 2 (default) or 3.
proj_features	bool	No	If True, draw biplot arrows showing feature contributions. Default False.
colors	list or tuple	No	Colors for each class or a single color for all points.
colormap	str or cmap	No	Matplotlib colormap for coloring points.
alpha	float	No	Transparency of scatter points. Default 0.75.
random_state	int, RandomState, or None	No	Random state for the PCA solver (used when randomized solver is enabled). Default None.
colorbar	bool	No	If True and target is continuous, draw a colorbar. Default True.
heatmap	bool	No	If True, draw a heatmap of feature contributions below the scatter plot. Not compatible with 3D. Default False.

Outputs

Name	Type	Description
pca_components_	ndarray, shape (n_components, n_features)	The principal component loadings matrix from the fitted PCA transformer.
classes_	ndarray, shape (n_classes,)	Class labels (discrete target only).
features_	ndarray, shape (n_features,)	Feature names discovered or provided during fit.
Xp (return from transform)	ndarray, shape (n, projection)	The PCA-transformed feature matrix.
ax	matplotlib Axes	The axes object containing the rendered scatter plot.

Usage Examples

Basic Usage

from yellowbrick.features import PCA as PCAVisualizer
from sklearn.datasets import load_iris

X, y = load_iris(return_X_y=True)

visualizer = PCAVisualizer(
    scale=True,
    projection=2,
    classes=["setosa", "versicolor", "virginica"],
    proj_features=True,
)
visualizer.fit(X, y)
visualizer.transform(X, y)
visualizer.show()

3D Projection

from yellowbrick.features import PCA as PCAVisualizer
from sklearn.datasets import load_iris
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

X, y = load_iris(return_X_y=True)

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

visualizer = PCAVisualizer(ax=ax, projection=3, colors=['r', 'g', 'b'])
visualizer.fit(X, y)
visualizer.transform(X, y)
visualizer.show()

Quick Method

from yellowbrick.features import pca_decomposition
from sklearn.datasets import load_iris

X, y = load_iris(return_X_y=True)

pca_decomposition(X, y, scale=True, projection=2, heatmap=True)

Related Pages

Implements Principle

• Principle:DistrictDataLabs_Yellowbrick_PCA_Projection

Requires Environment

• Environment:DistrictDataLabs_Yellowbrick_Python_Scikit_Learn_Environment