Disaster-Gpt

Frontpage
message

Inspiration

Our inspiration was to close the bridge between people in need and the instructions of what they need to do and combine it into a single solution Disaster-Gpt!

What it does

Disaster-Gpt receives the users information, of their name, location and phone number, then sending it to our database. Disaster-GPT then leverages the power of Gemini AI to continuously monitor real-time weather data and conduct sophisticated analysis every 6 hours. Using advanced machine learning models, Disaster-GPT can accurately detect patterns indicating an impending natural disaster and assess its severity. When a threat is identified, the AI not only predicts the exact timing and potential impact but also provides tailored, step-by-step guidance to users on how to stay safe. By combining cutting-edge AI with real-time data, Disaster-GPT ensures users receive fast, precise, and proactive support during critical moments

How we built it

We built the frontend using React with Vite, styled with Tailwind CSS The backend is powered by Express.js, and we integrated Tomorrow.io’s API to fetch real-time weather data, which we then process and wrangle using Python to clean up the data. which we analyze leveraging Google Gemini AI API for advanced insights. Finally, we keep users informed by sending SMS alerts through the Twilio API, ensuring they receive timely updates.

Challenges we ran into

A particular issue that we encountered throughout the project was finding the necessary weather data to feed into our genAI API. Our goal was to find an API that contained the required meteorological data so that we can leverage genAI to create predictive models on when a natural disaster is about to occur and use this information to inform our user base. Unfortunately the data that we found were either incomplete and did not contain the relevant information such as accurate coordinates or did not have access to past historical data or were locked behind a subscription service.

Accomplishments that we're proud of

An accomplishment that we as a team are proud of is being able to complete our project on time despite the tight time constraints. Especially because this is the first hackathon for most of our team members. The experience empowered each team member to take on new challenges, leading to a deeper understanding of various coding techniques and tools.

What we learned

We learned how to work with commonly used APIs like GenAI and Twilio, effectively leveraging their data and using HTTP protocols to manage communication between routes. This experience helped us understand how to integrate external services to enhance our application's functionality and responsiveness.

What's next for Disaster-Gpt

Like any AI model, Disaster-GPT would greatly benefit from access to more data to enhance its accuracy and predictive power. Additionally, learning how to effectively host the project would be a significant step in elevating its performance and accessibility, making it a more robust and reliable tool.