Customer Segmentation Project

Overview

This project implements end-to-end customer segmentation using RFM (Recency, Frequency, Monetary) analysis on retail transaction data. The system preprocesses the data, engineers features, applies clustering algorithms, and presents the results through an interactive Streamlit dashboard with actionable business insights.

Features

• Data preprocessing pipeline: Cleans and transforms retail transaction data
• RFM feature engineering: Creates and normalizes customer features based on purchase behavior
• Advanced customer segmentation: Uses K-means clustering with silhouette analysis
• Interactive Streamlit dashboard: Visualizes segments with 3D plots, radar charts, and more
• Business insights and recommendations: Provides actionable strategies for each customer segment
• ROI calculator: Estimates potential returns from segment-targeted campaigns

Project Structure

customer_segmentation_project/
│
├── data/                         # Data storage
│   ├── raw/                      # Original data files
│   │   └── online_retail_combined_dataset.csv
│   └── processed/                # Cleaned and processed data
│       ├── cleaned_retail_data.csv
│       ├── rfm_data.csv
│       ├── customer_features.csv
│       └── customer_segments.csv
│
├── notebooks/                    # Jupyter notebooks
│   └── exploratory_analysis.ipynb
│
├── src/                          # Source code
│   ├── data/
│   │   ├── data_loader.py        # Functions to load and validate data
│   │   └── data_processor.py     # Data cleaning and preprocessing
│   │
│   ├── features/
│   │   └── feature_engineering.py # RFM features creation
│   │
│   ├── models/
│   │   └── clustering.py         # KMeans and other clustering methods
│   │
│   └── visualization/
│       └── visualize.py          # Functions for visualization
│
├── streamlit/                    # Streamlit dashboard
│   ├── app.py                    # Main dashboard application
│   └── pages/
│       ├── 01_Data_Overview.py   # Data exploration
│       ├── 02_Customer_Segments.py # Segment analysis
│       └── 03_Business_Insights.py # Strategic recommendations
│
├── models/                       # Saved models and scalers
│   ├── kmeans_model.pkl
│   └── rfm_scaler.pkl
│
├── reports/                      # Generated analysis reports
│   └── figures/                  # Generated graphics and plots
│
├── requirements.txt              # Project dependencies
└── README.md                     # Project documentation

Installation

Prerequisites

• Python 3.8+
• pip

Setup

1. Clone the repository:

git clone https://github.com/yourusername/customer-segmentation-project.git
cd customer-segmentation-project

2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install the dependencies:

pip install -r requirements.txt

4. Place the raw data file in the data/raw/ directory.

Usage

Data Processing

# Run the data preprocessing pipeline
python -m src.data.data_processor

# Run the feature engineering pipeline
python -m src.features.feature_engineering

# Run the clustering pipeline
python -m src.models.clustering

Running the Dashboard

streamlit run streamlit/app.py

Dashboard Features

Main Dashboard

• Key metrics overview (total customers, average values)
• Customer segment distribution
• 3D cluster visualization
• RFM metrics distribution

Data Overview

• Data quality assessment
• Time series analysis of sales
• Geographic distribution of customers
• Product analysis and trends

Customer Segments

• Detailed segment profiles
• Interactive comparisons between segments
• Segment characteristic visualization
• Customer explorer for individual analysis

Business Insights

• Revenue opportunity identification
• Segment-specific strategic recommendations
• Campaign planning and calendar
• ROI calculator for marketing initiatives

Customer Segments

The project identifies distinct customer segments:

Segment	Description	Recommended Actions
Champions	High-value, frequent, recent customers	Reward loyalty, encourage ambassadorship
Loyal Customers	Regular, active customers	Increase purchase value, special offers
Big Spenders	High-value but infrequent	Increase purchase frequency, personalized communication
New Customers	Recent first-time buyers	Convert to repeat customers, second purchase incentives
At Risk	Previously active, becoming inactive	Reactivation campaigns, special offers
Dormant	Long-inactive customers	Re-engagement, win-back promotions
Rare Shoppers	Infrequent, low-value	Increase purchase frequency and value

Model Performance

• The clustering model uses silhouette score and inertia to determine optimal number of clusters
• Customer segment labels are validated through business rules and domain knowledge
• Segment recommendations are based on RFM metrics and retail industry best practices

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• The dataset used is the Online Retail dataset from the UCI Machine Learning Repository
• Thanks to all contributors and the open-source community for their valuable tools and libraries