🛍️ Online Retail Customer Value Analysis System

📋 Project Overview

This is a complete AI/Data Analysis project for analyzing online retail data, identifying high-value customers, and building customer segmentation models. The project demonstrates a complete workflow from data understanding, processing, model building to validation and evaluation.

⭐ Core Skills Demonstrated

• ✅ Python Data Processing: pandas, numpy for data cleaning and feature engineering
• ✅ Reproducible Scripts: Modular design with clear code structure
• ✅ Testing & Validation: Defined inputs/outputs, edge cases, and evaluation metrics
• ✅ Comparative Evaluation: Baseline vs improved version performance comparison
• ✅ Engineering Practices: Git version control, clear documentation, runnable project structure

📁 Project Structure

.
├── README.md                 # Project documentation
├── requirements.txt          # Dependencies list
├── .gitignore               # Git ignore file
├── data/
│   └── online_retail_II.csv # Raw data
├── src/
│   ├── data_preprocessing.py    # Data preprocessing module
│   ├── rfm_analysis.py          # RFM customer value analysis
│   ├── customer_segmentation.py # Customer segmentation model
│   └── evaluation.py            # Evaluation and validation module
├── tests/
│   ├── test_preprocessing.py    # Preprocessing tests
│   ├── test_rfm.py              # RFM analysis tests
│   └── test_evaluation.py       # Evaluation tests
├── notebooks/
│   └── customer_analysis.ipynb  # Complete analysis workflow notebook
└── results/
    └── README.md                # Results documentation

🚀 Quick Start

1. ⚙️ Environment Setup

# Create virtual environment (recommended)
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

2. ▶️ Run Analysis

Option 1: Using Jupyter Notebook (Recommended)

jupyter notebook notebooks/customer_analysis.ipynb

Option 2: Run Python Scripts

# Data preprocessing
python src/data_preprocessing.py

# RFM analysis
python src/rfm_analysis.py

# Customer segmentation
python src/customer_segmentation.py

# Evaluation and validation
python src/evaluation.py

3. 🧪 Run Tests

pytest tests/

🧩 Project Components

1. 🧹 Data Preprocessing (src/data_preprocessing.py)

• Data cleaning: Handle missing values, outliers
• Data validation: Check data quality and completeness
• Feature engineering: Extract time features, calculate derived metrics

Input: Raw CSV file Output: Cleaned DataFrame Validation Metrics: Data completeness, missing rate, outlier detection

2. 📊 RFM Analysis (src/rfm_analysis.py)

• Recency: Most recent purchase time
• Frequency: Purchase frequency
• Monetary: Spending amount
• Customer value scoring

Input: Cleaned transaction data Output: RFM scores and customer segments Evaluation Metrics: Segment rationality, customer distribution uniformity

3. 👥 Customer Segmentation (src/customer_segmentation.py)

• Baseline method: Rule-based segmentation using RFM
• Improved method: K-means clustering + RFM features
• Performance comparison evaluation

Input: RFM feature data Output: Customer segmentation results and evaluation report Evaluation Metrics:

• Silhouette Score
• Intra-cluster cohesion
• Inter-cluster separation
• Business metrics (average value per segment)

4. ✅ Evaluation & Validation (src/evaluation.py)

• Define test cases and edge cases
• Calculate evaluation metrics (accuracy, precision, recall, etc.)
• Generate comparison reports

Test Scenarios:

• Normal data flow
• Edge cases (single-purchase customers, abnormal large orders, etc.)
• Data quality validation

📈 Evaluation Metrics

🔍 Data Quality Metrics

• Data completeness: > 95%
• Missing value rate: < 5%
• Outlier detection rate

⚡ Model Performance Metrics

• Baseline Method:

  • Number of segments: 8 (based on RFM rules)
  • Computation speed: Fast
  • Interpretability: High

• Improved Method:

  • Silhouette Score: > 0.3
  • Intra-cluster cohesion: 20%+ improvement
  • Inter-cluster separation: 15%+ improvement

💰 Business Metrics

• High-value customer identification accuracy
• Customer churn prediction accuracy
• Average spending per segment

🛠️ Tech Stack

• Python 3.8+
• pandas: Data processing
• numpy: Numerical computation
• scikit-learn: Machine learning models
• matplotlib/seaborn: Data visualization
• pytest: Unit testing
• jupyter: Interactive analysis

✨ Project Highlights

1. Complete Engineering Practices: Modular design, test coverage, comprehensive documentation
2. Verifiable Components: Each module has clear inputs/outputs and validation metrics
3. Comparative Evaluation: Quantitative comparison of Baseline vs improved versions
4. Edge Case Handling: Processing of abnormal data and extreme cases
5. Reproducibility: Clear code structure and dependency management

📄 License

MIT License