Implementation:DistrictDataLabs Yellowbrick TSNEVisualizer

Knowledge Sources	DistrictDataLabs_Yellowbrick
Domains	NLP, Visualization, Dimensionality_Reduction
Last Updated	2026-02-08 05:00 GMT

Overview

Concrete tool for visualizing document similarity in 2D space using t-SNE dimensionality reduction, provided by the Yellowbrick text module.

Description

The TSNEVisualizer applies t-SNE (t-distributed Stochastic Neighbor Embedding) to high-dimensional document vectors to produce a 2D scatter plot where similar documents cluster together. It supports optional preliminary dimensionality reduction via SVD or PCA before applying t-SNE, and colors points by document class labels.

Usage

Import this visualizer when exploring document similarity in a text corpus. It works with pre-vectorized document-term matrices and is useful for identifying clusters of similar documents.

Code Reference

Source Location

• Repository: DistrictDataLabs_Yellowbrick
• File: yellowbrick/text/tsne.py
• Lines: 1-428

Signature

class TSNEVisualizer(TextVisualizer):
    def __init__(
        self,
        ax=None,
        decompose="svd",
        decompose_by=50,
        labels=None,
        classes=None,
        colors=None,
        colormap=None,
        random_state=None,
        alpha=0.7,
        **kwargs,
    ):
        """t-SNE document similarity visualizer."""

    def make_transformer(self, decompose="svd", decompose_by=50, tsne_kwargs={}):
        """Creates the decomposition + t-SNE pipeline."""

def tsne(
    X, y=None, ax=None, decompose="svd", decompose_by=50, labels=None,
    colors=None, colormap=None, alpha=0.7, show=True, **kwargs,
):
    """Quick method for one-off t-SNE visualization."""

Import

from yellowbrick.text import TSNEVisualizer
from yellowbrick.text.tsne import tsne

I/O Contract

Inputs

Name	Type	Required	Description
X	sparse or dense matrix	Yes	Document-term matrix (fit)
y	array-like	No	Document labels for coloring
decompose	str	No	Pre-reduction: "svd" or "pca" (default: "svd")
decompose_by	int	No	Intermediate dimensions (default: 50)
alpha	float	No	Point transparency (default: 0.7)

Outputs

Name	Type	Description
ax	matplotlib.Axes	Axes with 2D t-SNE scatter plot

Usage Examples

from sklearn.feature_extraction.text import TfidfVectorizer
from yellowbrick.text import TSNEVisualizer
from yellowbrick.datasets import load_hobbies

corpus = load_hobbies()
tfidf = TfidfVectorizer()
X = tfidf.fit_transform(corpus.data)

viz = TSNEVisualizer()
viz.fit(X, corpus.target)
viz.show()

Related Pages

• Environment:DistrictDataLabs_Yellowbrick_Python_Scikit_Learn_Environment