🛍️ Customer Segmentation using K-Means Clustering

📌 Project Overview

This project applies Unsupervised Machine Learning to segment retail customers based on their purchasing behavior. By analyzing annual income and spending patterns, I identified five distinct customer "personas," enabling businesses to transition from mass marketing to highly personalized, data-driven strategies.

🛠️ Tech Stack

• Language: Python
• Libraries: Scikit-Learn, Pandas, Matplotlib, Seaborn
• Algorithm: K-Means Clustering

🔍 Technical Workflow

• Data Source: Mall Customer Segmentation dataset (Local .csv).
• Feature Selection: Focused on Annual Income (k$) and Spending Score (1-100) to map behavioral segments.
• Optimization: Implemented the k-means++ initialization method to ensure faster convergence and avoid suboptimal clustering.

---

📊 Visual Analysis (Click to Expand)

1. Determining Optimal Clusters (The Elbow Method)

To find the mathematical "sweet spot" for the number of clusters, I calculated the Within-Cluster Sum of Squares (WCSS) for 1 to 10 clusters.

Analysis: The "elbow" clearly forms at k=5, indicating that adding more clusters beyond this point provides diminishing returns in explaining data variance.

2. Customer Persona Visualization

The final model successfully grouped 200 customers into five behavioral segments.

Strategic Personas Identified:

• 🔴 Sensible: High Income, Low Spending (Target for high-value savings/investments).
• 🔵 Standard: Average Income, Average Spending (The stable core customer base).
• 🟢 Target/Whales: High Income, High Spending (Primary targets for luxury loyalty programs).
• 🟡 Careless: Low Income, High Spending (High-engagement, impulse buyers).
• 🟣 Miser: Low Income, Low Spending (Highly price-sensitive shoppers).

---