This is the Analytics Service component of the Splitwise Clone project, implementing 4 core responsibilities:

1. Spending pattern analysis and visualization
2. Group spending statistics aggregation
3. ML-based expense categorization
4. Historical trend analysis

split_data/
├── Core Python Files
│   ├── analysis.py              # Spending patterns, group stats, trend analysis
│   ├── chart_data.py            # Chart data generation for visualization
│   ├── expense_classifier.py   # ML-based expense categorization
│   ├── database.py              # Database connection and queries
│   └── api_server.py            # Flask API server
│
├── ML Components
│   ├── train_classifier.py      # Train the ML model
│   └── expense_classifier_model.pkl  # Trained model (run train_classifier.py first)
│
├── Database
│   └── init.sql                 # Database schema
│
├── Docker
│   ├── Dockerfile               # Container configuration
│   └── docker-compose.yml      # Full stack setup (API + MySQL)
│
├── Testing
│   ├── test_all_responsibilities.py  # Test all 4 responsibilities
│   ├── test_classifier.py      # Test ML classifier
│   ├── test_workflow.py        # Test complete workflow
│   └── test_settlement_logic.py  # Test settlement logic
│
├── Utilities
│   ├── generate_dummy_data.py  # Generate test data
│   ├── settlement_checker.py   # Background settlement checker service
│   ├── chart_data_api.py       # CLI script for chart data
│   └── api_example.py          # API usage examples
│
└── Documentation
    ├── RESPONSIBILITIES_GUIDE.md  # Complete API documentation
    ├── HOW_TO_TEST.md            # Testing guide
    └── DOCKER_SETUP.md           # Docker setup guide

• Python 3.8+
• MySQL (via Docker)
• Docker & Docker Compose

1. Install dependencies:

   pip install -r requirements.txt

2. Train the ML model (first time only):

   python3 train_classifier.py

3. Start the database:

   docker-compose up -d mysql

4. Generate test data (optional):

   python3 generate_dummy_data.py

5. Start the API server:

   python3 api_server.py

   Or use Docker Compose for everything:

   docker-compose up -d

• GET /api/users/<user_id>/analysis/patterns - Spending pattern analysis
• GET /api/groups/<group_id>/statistics - Group spending statistics
• POST /api/tags/suggest - ML-based expense categorization (top 3 suggestions)
• GET /api/users/<user_id>/analysis/trends - Historical trend analysis

• GET /api/users/<user_id>/charts - All chart data
• GET /api/users/<user_id>/charts/weekly - Weekly expenses
• GET /api/users/<user_id>/charts/monthly - Monthly expenses
• GET /api/users/<user_id>/charts/categories - Expenses by category

See RESPONSIBILITIES_GUIDE.md for complete API documentation.

Run all tests:

python3 test_all_responsibilities.py

Test individual components:

python3 test_classifier.py      # Test ML classifier
python3 analysis.py             # Test analysis functions
python3 test_workflow.py        # Test complete workflow

See HOW_TO_TEST.md for detailed testing instructions.

• Language: Python 3.8+
• Framework: Flask (REST API)
• ML Library: scikit-learn (TF-IDF + Naive Bayes)
• Database: MySQL
• Containerization: Docker & Docker Compose

• RESPONSIBILITIES_GUIDE.md - Complete API documentation with examples
• HOW_TO_TEST.md - Testing guide for all 4 responsibilities
• DOCKER_SETUP.md - Docker deployment guide

Jiawei Li - Analytics Service Implementation