AI Bank Statement Analyzer

An intelligent bank statement analysis tool that leverages artificial intelligence to automatically process, categorize, and analyze financial transactions from bank statements. This application provides insights into spending patterns, transaction categorization, and financial analytics.

UI Demo link

https://t.co/vo9CV47OjF

ERD diagram

Features

• Automated Transaction Processing: Upload and process bank statements in various formats
• AI-Powered Categorization: Intelligent categorization of transactions using machine learning
• Financial Analytics: Generate comprehensive reports and insights from transaction data
• Asynchronous Processing: Background task processing using Celery for handling large datasets
• Real-time Monitoring: Built-in task monitoring with Celery Flower
• RESTful API: Clean API endpoints for integration with other applications
• Data Visualization: Interactive charts and graphs for financial insights

Technology Stack

• Backend: Python with Flask/FastAPI
• Task Queue: Celery with Redis/RabbitMQ
• AI/ML: Machine learning models for transaction analysis and categorization
• Database: PostgreSQL/SQLite for data persistence
• Monitoring: Celery Flower for task monitoring
• Containerization: Docker support (optional)

Prerequisites

• Python 3.8+
• Redis or RabbitMQ (for Celery message broker)
• PostgreSQL (if using PostgreSQL as database)

Installation

Local Setup

1. Clone the repository

   git clone <repository-url>
   cd AI-bank-statement-analyzer

2. Create virtual environment

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

3. Install dependencies

   pip install -r requirements.txt

4. Set up environment variables

   cp .env.example .env
   # Edit .env file with your configuration

5. Initialize database

   alembic upgrade head

   if a change is made to the DB then run this command to perform the migration(s)

   alembic revision --autogenerate -m "migration_name"
   alembic upgrade head

Docker Setup (Optional)

You can also run the application using Docker for easier deployment and development:

1. Build and run with Docker Compose

   docker-compose up --build

2. Or build individual containers

   docker build -t ai-bank-analyzer .
   docker run -p 5000:5000 ai-bank-analyzer

Running the Application

1. Start the Message Broker

Ensure Redis or RabbitMQ is running on your system.

For Redis:

redis-server

2. Start the Celery Worker

celery -A app.tasks.celery_app worker -l info

3. Start Celery Flower (Optional - for monitoring)

celery -A app.tasks.celery_app flower --port=5555

4. Start the Main Application

python app.py

The application will be available at:

• Main Application: http://localhost:8000
• Celery Flower Dashboard: http://localhost:5555

API Endpoints

Upload Bank Statement

POST /api/upload
Content-Type: multipart/form-data

Get Analysis Results

GET /api/analysis/{analysis_id}

Get Transaction Categories

GET /api/categories

Get Financial Summary

GET /api/summary/{period}

Usage

1. Upload Bank Statement: Use the web interface or API to upload your bank statement (PDF, CSV, or Excel format)

2. Processing: The system will automatically:

   • Extract transaction data
   • Clean and normalize the data
   • Apply AI models for categorization
   • Generate insights and analytics

3. View Results: Access your analysis through:

   • Web dashboard for visual insights
   • API endpoints for programmatic access
   • Export options for further analysis

Configuration

Key configuration options in your .env file:

# Database
DATABASE_URL=postgresql://user:password@localhost/dbname

# Celery
CELERY_BROKER_URL=redis://localhost:6379/0
CELERY_RESULT_BACKEND=redis://localhost:6379/0

# AI Models
OPENAI_API_KEY=your_openai_key_here

# Application
FLASK_ENV=development
SECRET_KEY=your_secret_key_here

Monitoring

Celery Flower Dashboard

Access the Celery Flower dashboard at http://localhost:5555 to monitor:

• Active tasks
• Task history
• Worker status
• Task execution times
• Failed tasks and error logs

Application Logs

Monitor application logs for debugging and performance tracking:

tail -f logs/app.log

Supported File Formats

• PDF: Bank statements in PDF format
• CSV: Comma-separated transaction files
• Excel: .xlsx and .xls files

AI Features

• Transaction Categorization: Automatic categorization into predefined categories (Food, Transportation, Entertainment, etc.)
• Merchant Recognition: Intelligent merchant name normalization
• Spending Pattern Analysis: Identification of spending trends and patterns
• Anomaly Detection: Detection of unusual transactions
• Budget Recommendations: AI-powered budget suggestions based on spending history

Troubleshooting

Common Issues

1. Celery Worker Not Starting

   • Ensure message broker (Redis/RabbitMQ) is running
   • Check celery configuration in settings

2. File Upload Errors

   • Verify file format is supported
   • Check file size limits
   • Ensure proper permissions

3. Database Connection Issues

   • Verify database credentials
   • Ensure database server is running
   • Check network connectivity

Debug Mode

Run the application in debug mode for detailed error messages:

export FLASK_ENV=development
python app.py

Performance Optimization

• Batch Processing: Large files are processed in chunks
• Caching: Redis caching for frequently accessed data
• Database Indexing: Optimized database queries
• Async Processing: Background processing for time-intensive tasks