post-analysis-clustering

A Python package for visualizing and interpreting clustering results using statistical tests, feature importance, and insightful plots.

📦 Installation

Install via pip:

pip install post-analysis-clustering

📦 post-analysis-clustering

📊 Overview

The post-analysis-clustering package is designed to help analyze, validate, and interpret clustering results. It provides tools to:

• Visualize feature distributions across clusters
• Identify distinguishing features using statistical tests
• Plot heatmaps, snake plots, importance charts, and more
• Evaluate inter-cluster separation and overlap

It is especially useful for clustering results from customer segmentation, fraud detection, or other unsupervised learning pipelines.

🔧 Features

• 📈 Box, Violin, and Distribution Plots for feature-by-cluster analysis
• 🧮 Permutation Importance Heatmaps across multiple classifiers
• 📊 Crosstab and Binned Heatmaps to explore categorical and continuous variables
• ✅ Chi-square tests with human-readable significance interpretation
• 🎨 Custom color palettes for consistent cluster visualization

🚀 Usage

Basic usage example:

from post_analysis_clustering import plot_bin_heatmap

plot_bin_heatmap(
    raw_df=df,
    features=["age", "income", "purchase_amount"],
    target_cluster="cluster",
    annot_type="Percentage"
)

For a complete usage example, check out the dev.ipynb.

👤 Author

• 🔗 GitHub Profile
• 💼 LinkedIn