Credit Card Customer Segmentation

Project Overview

This project focuses on segmenting credit card customers based on their spending behavior and demographics. Using clustering techniques such as K-means, Agglomerative Clustering, and Gaussian Mixture Models, the analysis aims to uncover distinct customer groups. The insights generated help support targeted marketing strategies and personalized customer experiences.

Dataset

Dataset Source: Credit Card Dataset

Name	Description
custID	Identification of Credit Card holder (Categorical)
remainBal	Balance amount left in their account to make purchases
balF	How frequently the Balance is updated, score between 0 and 1 (1 = frequently updated, 0 = not frequently updated)
purDone	Amount of purchases made from account
oneOffPur	Maximum purchase amount done in one-go
insPur	Amount of purchase done in installment
cashAdv	Cash in advance given by the user
purF	How frequently the Purchases are being made, score between 0 and 1 (1 = frequently purchased, 0 = not frequently purchased)
oneOffPurF	How frequently Purchases are happening in one-go (1 = frequently purchased, 0 = not frequently purchased)
insPurF	How frequently purchases in installments are being done (1 = frequently done, 0 = not frequently done)
cashAdvF	How frequently the cash in advance being paid
cashAdvTRX	Number of Transactions made with "Cash in Advanced"
purTRX	Numbe of purchase transactions made
creditLim	Limit of Credit Card for user
pymtDone	Amount of Payment done by user
minPymt	Minimum amount of payments made by user
fullPymtPCT	Percentage of full payment paid by user
tenure	Tenure of credit card service for user

Project Objectives

1. Data Cleaning & Preprocessing

   • Rename inconsistent feature names for clarity
   • Handle missing values using cold deck imputation (mean substitution)
   • Detect and remove outliers using interquartile range (IQR) and Z-score analysis

2. Exploratory Data Analysis (EDA)

   • Compute key statistical summaries: mean, median, variance, and standard deviation
   • Perform correlation analysis using heatmaps
   • Visualize distributions with histograms and scatter plots
   • Detect outliers using boxplots

3. Machine Learning Models for Customer Segmentation

   • Data Normalization: Apply Standard Scaler to standardize feature distributions
   • Clustering Algorithms: Implement and compare
     • K-means
     • Agglomerative Clustering
     • Gaussian Mixture Model (GMM)
   • Dimensionality Reduction: Use Principal Component Analysis (PCA) for feature compression
   • Cluster Validation & Model Evaluation:
     • Elbow Method: Determine optimal cluster count
     • Silhouette Score: Evaluate cluster compactness
     • Davies-Bouldin Index & Calinski-Harabasz Method: Assess clustering performance
   • Predictive Modeling: Assign new customers to clusters based on trained segmentation model

Technologies Used

• Programming Language: Python
• Libraries: pandas, numpy, scikit-learn, matplotlib
• Data Visualization Tools: seaborn, plotly

Project Workflow

1. Data Collection: Import and inspect datasets
2. Data Cleaning & Preprocessing: Handle missing values, normalize data, and remove outliers
3. Exploratory Data Analysis (EDA): Visualize distributions, correlations, and patterns
4. Feature Engineering: Transform variables for better clustering
5. Model Training: Implement and compare different clustering algorithms
6. Model Evaluation: Analyze clustering performance using multiple validation metrics
7. Results Interpretation: Identify customer segments and derive actionable insights