Hurristics

Inspiration

My team and I come from South Florida, a location constantly bombarded and worried about dealing with hurricanes and the precautions that come alongside them. We wish to have a system that could accurately predict the trajectory of a hurricane using deep learning and provide data to help agencies decide where resources need to be for proper allocation.

What it does

Hurristics is a prediction model that provides a hurricane path and a heatmap indicating the risks and damage associated with the distance of the hurricane and its strength.

How we built it

Our backend is built on Python and Flask, Microsoft Azure, GCP, and Javascript. We used Jupyter for coming up with the prediction of the hurricane model. Our frontend and infrastructure was built around React.

Challenges we ran into

One of the biggest issues we came across was representing longitude and latitude in an easy way for our frontend API to understand and pull from. While learning Mapbox and it's functions, we struggled with displaying the path of the hurricane, although we are able to predict it.

Accomplishments that we're proud of

We got to utilize new technologies outside of our comfort zone and make a fully functional prediction model utilizing deep learning AI. Our backend operations were all fully operational but we ended up struggling with representing our data on the front end.

What we learned

For this project, every member of the team utilized new technologies we're unfamiliar with to produce an optimized and effective model for hurricane predictions and representation.

What's next for Hurristics

After HackGT, we wish to implement a fully functional front-end that is more fleshed out and provides more features. We also wish to make our API more flexible for others to use.

Built With

• appsurface
• azure
• firebase
• gcp
• javascript
• jupyter
• mapbox
• python
• react
• typescript