AI Hackfest 2023 - Receipt Analyzer

Note: This project is still in progress and is currently at the MVP stage.

Description

The AI Hackfest 2023 Receipt Analyzer is a Flask web application developed as part of the AI Hackfest organized by Major League Hackers (MLH). The goal of this project is to create a receipt scanning and categorization tool using Optical Character Recognition (OCR), a Large Language Model (GPT 3.5), and a MongoDB database for data storage. Additionally, it provides analytics capabilities using the Pandas library in Python.

Technologies Used

The following technologies were used to develop the AI Hackfest 2023 Receipt Analyzer:

• Python
• HTML & CSS
• Flask
• EasyOCR
• GPT (Large Language Model)
• MongoDB

Dependencies

Make sure the following dependencies are installed:

• Python
• Flask
• EasyOCR
• GPT
• MongoDB

Features

The Receipt Analyzer web application offers the following features:

1. Receipt Scanning: Users can upload images of receipts, which are then processed using Optical Character Recognition (OCR) to extract text from the images.
2. Purchase Categorization: The extracted text is then passed through a Large Language Model (GPT 3.5) to categorize the purchases mentioned in the receipt.
3. Data Storage: The categorized purchase data is stored in a MongoDB database for future retrieval and analysis.
4. Analytics: The application provides analytics capabilities using the Pandas library in Python. Users can generate reports, perform data analysis, and visualize the purchase data.

Installation and Usage

To run the AI Hackfest 2023 Receipt Analyzer locally, please follow these steps:

1. Clone the repository: git clone https://github.com/shayaansultan/ai-hackfest-2023.git
2. Navigate to the project directory: cd ai-hackfest-2023
3. Install the required dependencies mentioned above.
4. Set up a MongoDB database and update the database configuration in config.py.
5. Run the Flask application: python app.py
6. Access the web application in your browser at http://localhost:5000.

Future Enhancements

While the AI Hackfest 2023 Receipt Analyzer is currently at the MVP stage, there are several areas for future enhancement:

• Improved OCR Accuracy: Enhancing the OCR process to improve accuracy and handle various types of receipts.
• Refining Categorization: Further training the Large Language Model (GPT 3.5) to better categorize purchases and handle a wider range of receipt formats.
• User Authentication: Implementing user authentication and authorization to secure the application and enable personalized experiences.
• Advanced Analytics: Adding advanced data visualization and analytics features to provide deeper insights into purchase patterns and trends.
• Deployment: Deploying the application to a cloud hosting platform to make it accessible from anywhere.

Contributing

Contributions to the AI Hackfest 2023 Receipt Analyzer project are welcome. If you have any ideas, suggestions, or bug reports, please feel free to open an issue or submit a pull request.

License

The AI Hackfest 2023 Receipt Analyzer project is open-source and available under the MIT License.

Acknowledgements

We would like to thank Major League Hackers (MLH) for organizing the AI Hackfest 2023 and providing the opportunity to develop this project. I would also like to thank @YuvBindal for his contributions to this project.

Special thanks to the creators and maintainers of the following libraries used in this project:

• Flask: https://flask.palletsprojects.com/
• EasyOCR: https://github.com/JaidedAI/EasyOCR
• GPT: https://github.com/openai/gpt-3.5