SQL & K-Means

Project 3 – Customer Segmentation with SQL & K-Means Clustering

Problem:
A business needed to segment its customers to improve marketing strategies, personalize offers, and boost retention. Traditional demographic segmentation was not delivering actionable insights.

Approach:
I combined SQL for data preprocessing and K-Means clustering in Python to segment customers based on purchasing behavior.

• Used SQL (via DuckDB) to clean and explore the raw dataset.
  
  

• Selected relevant features such as Annual Income and Spending Score.
  
  

• Scaled the data and applied K-Means Clustering to identify meaningful customer segments.
  
  

• Visualized clusters using matplotlib to show distinct customer groups.
  
  

Tools:

• DuckDB: SQL queries within Python
  
  

• pandas: Data handling
  
  

• scikit-learn: K-Means clustering
  
  

• matplotlib, seaborn: Visualization
  
  

Outcome:
The analysis uncovered three clear customer segments:

• High-income, high-spending customers (ideal for loyalty programs),
  
  

• Low-income, high-spending (price-sensitive but brand-loyal),
  
  

• High-income, low-spending (potential for upselling).
  
  

These insights can drive targeted marketing, improve customer retention, and increase ROI on campaigns.