Product Recommendation System & Behavioral Insights Engine

This project explores customer segmentation and product recommendation using real-world e-commerce transactional data. By applying RFM clustering and association rule mining, the system provides data-driven insights into purchasing behavior and recommends relevant products to boost engagement and retention. The end-to-end pipeline includes data wrangling, behavioral analysis, recommendation generation, and an interactive Streamlit dashboard to simulate customer and product behavior.

Python (Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn),
Machine Learning (K-Means Clustering),
Market Basket Analysis (Apriori, Association Rules),
Streamlit (Dashboard),
Data Source: Online Retail Dataset (UCI Repository)
Version Control: Git & GitHub

Tools & Technologies Used

Cleaned and transformed raw e-commerce transactional data to prepare it for analysis.
Engineered features like Total Revenue and used RFM (Recency, Frequency, Monetary) metrics to segment customers.
Applied K-Means clustering to identify meaningful customer groups based on purchasing behavior.
Built an association rules–based recommendation engine using the Apriori algorithm to discover high-lift product pairs.
Designed and deployed a responsive Streamlit dashboard to visualize customer segments, rule-based recommendations, and key behavioral metrics.
Documented the entire project pipeline and optimized code for GitHub presentation and recruiter viewing.

What I did

Key Highlights

• RFM-based Customer Segmentation
• K-Means Clustering Visualization
• Product Recommendation using Association Rules
• Behavioral Analytics Dashboard (Streamlit)
• Cleaned, processed, and reusable dataset with deployable model

Links

Github: Product Recommendation System & Behavioral Insights Engine

Live App: product-recommendation-otimisation-app