🛍️ Customer Behavior Clustering with Categorical Features

This project explores customer behavior by applying clustering techniques on a dataset with categorical and numerical features. The analysis involves preprocessing, statistical testing, dimensionality reduction, and clustering to uncover hidden patterns in customer preferences.

📌 Objectives

• Cluster customers based on their behavior and attributes.
• Apply dimensionality reduction using PCA for visualization.
• Use statistical tests to explore feature relationships.

🔍 Features Used

The following features were selected for clustering:

• Gender
• Item Purchased
• Location
• Shipping Type
• Payment Method
• Frequency of Purchases
• Subscription Status
• Age
• Purchase Amount (USD)
• Size (ordinal: S, M, L, XL)

Encoding Strategy

• Size was encoded using OrdinalEncoder with custom order.
• All other categorical features were encoded using OneHotEncoder.

🧪 Statistical Analysis

The notebook includes several statistical tests to understand relationships between features:

• Spearman Correlation for detecting associations between ordinal/numerical features.
• Chi-Square Test of Independence to check dependence between categorical variables.
• Cramér’s V to measure the strength of association between categorical features.

These tests guide preprocessing and support interpretation of clustering results.

⚙️ Clustering Workflow

Dimensionality Reduction

• PCA was applied after encoding to reduce high-dimensional data.
• The number of components was selected based on the explained variance ratio.

Clustering Algorithms

• KMeans Clustering was used with the optimal number of clusters estimated using the Elbow Method.
• Agglomerative Clustering was also used to compare hierarchical clustering performance.

📊 Visualization

• 2D and 3D PCA scatter plots were created using plotly to visualize clustering results.
• Plots help evaluate cluster separation and structure.

🎯 Marketing Strategy Implications

Cluster insights can support:

• Targeted marketing based on customer behavior.
• Improved segmentation for promotions or subscription offers.
• Data-driven decisions on product, shipping, and payment preferences.