gioche6 / flask-explainability-app

Creating an AI explainabilty Score App

Explainability Scoring System

This project is a web application that evaluates AI systems based on various criteria such as fairness, accountability, and transparency. The scores are visualized using a radial graph, and the data is stored in an SQLite database.

Features

• Fairness Assessment: Evaluates AI systems based on demographic parity, equal opportunity, and absence of disparate impact.
• Accountability Assessment: Scores AI systems on documentation practices, auditability, and compliance with ethical guidelines.
• Transparency Assessment: Measures explainability, user interface design, and decision-making documentation.
• Data Storage: Stores the assessment scores and generated graphs in an SQLite database.
• Graph Visualization: Displays a radial graph representing the scores.

Technologies Used

• Python: Programming language used for the backend.
• Flask: Micro web framework for Python.
• SQLite: Database for storing scores and graphs.
• Plotly: Library for creating interactive graphs.
• Bootstrap: Frontend framework for styling the web pages.
• Heroku: Platform for deploying the web application.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/explainability-scoring-system.git
   cd explainability-scoring-system

2. Create a virtual environment:

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

3. Install dependencies:

   pip install -r requirements.txt

4. Initialize the database:

   python database_setup.py

5. Run the application:

   python app.py

6. Open your browser and visit:

   http://127.0.0.1:5000
   

Deployment

To deploy the application on Heroku, follow these steps:

1. Login to Heroku:

   heroku login

2. Create a new Heroku app:

   heroku create your-app-name

3. Push the code to Heroku:

   git push heroku main

4. Run database setup on Heroku:

   heroku run python database_setup.py

5. Open the deployed app:

   heroku open

Usage

1. Home Page: Enter the AI system title and scores for each criterion, then click "Calculate".
2. View Results: Click "View Stored Scores" to see all stored scores and view individual score details and graphs.

Contributing

Contributions are welcome! Please fork the repository and create a pull request with your changes.

License

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

Contact

For any questions or feedback, please contact [email protected].