Overview

This project demonstrates the use of Principal Component Analysis (PCA) on a real-world credit card customer dataset to reduce dimensionality and analyze data structure.

What I Did

Cleaned and preprocessed the dataset

Removed identifier and target columns before applying PCA

Encoded categorical features and scaled numerical data

Applied PCA for:

Visualization (2 components)

Modeling (multiple components)

Evaluated the impact of PCA on classification performance

Key Learnings

PCA is an unsupervised technique used for dimensionality reduction

Fewer components help visualization, more components improve model performance

Explained variance is critical for selecting the number of components

PCA should not use the target variable to avoid data leakage

Tools Used

Python

Pandas, NumPy

Matplotlib

Scikit-learn

Files

PCA_Credit_Card_Analysis.ipynb – complete analysis and implementation