FloodGPT-4: An Advanced AI Assistant Enabled by GPT-4 for Enhanced Interpretability and Public Engagement
Welcome to the FloodGPT-4 project repository! This initiative explores the integration of GPT-4's advanced capabilities with flood risk management strategies to offer an innovative AI Assistant. Our prototype, featured in the paper "Towards Democratized Flood Risk Management: An Advanced AI Assistant Enabled by GPT-4 for Enhanced Interpretability and Public Engagement," aims to make flood risk information more accessible and actionable for the public.
This repository primarily contains the Jupyter Notebook (FloodGPT.ipynb
) that serves as the core of our prototype, showcasing the AI Assistant designed to enhance flood risk management.
To get the prototype up and running, follow these steps:
-
Clone this repository to your local machine to get the
FloodGPT.ipynb
notebook. You can do this by running the following command in your terminal or command prompt:git clone https://github.com/RafaelaMartelo/FloodGPT-4_Prototype.git cd FloodGPT-4_Prototype
-
Download the 2020 US SVI data from CDC/ATSDR's official site. This data is essential for the prototype's operation.
-
Place the downloaded SVI_2020_US folder in the same directory as the FloodGPT.ipynb notebook.
-
Create a .env file in the root directory of the cloned repository. Inside this file, include your API keys for the National Flood Map Data API and the ChatGPT-4 API like so:
OPENAI_API_KEY=your_openai_api_key_here FEMA_API_KEY=your_fema_api_key_here
-
Ensure you have access to the following APIs used in the prototype:
- National Flood Map Data API: Documentation - Used to retrieve flood data and flood map.
- ChatGPT-4 API: Documentation - Powers the intelligent responses and function calls within the prototype.
-
Open and run the notebook in a Jupyter environment. This will activate the FloodGPT-4 AI Assistant, ready to assist with flood risk management tasks.
Rafaela Martelo & Ruo-Qian Wang
This project's initial structure and function calling implementation were inspired by the following resources:
-
"Function Calling: Integrate Your GPT Chatbot With Anything" by Tomas Fernandez Link to Article
-
"gpt-function-calling-tutorial" by JayZeeDesign Link to GitHub Repository
This project is under the MIT License.