FloodSight

Team Members

Grace Mihiral is a Computer Engineering Major with the CISE and ECE Departments. Matas Ambrukaitis is a Computer Science Major, minoring in Digital Arts and Sciences, with the CISE Department and Digital Worlds Institute. Sophie Alfonso is a Mathematics and Computer Science Major with the Mathematics and CISE Departments.

Inspiration

Roughly 80% of Floridians live or work near the coast, making millions of people susceptible to severe flooding and storm surge. Yet, while we constantly hear warnings of 5 to 20 feet of flooding, most of us have never actually seen what that looks like in our own neighborhoods. We built FloodSight to change that by helping users identify flood zones in their area as well and visualize real-world water rise right in front of them.

What it does

FloodSight combines real-world elevation, soil drainage, historical flood records, and water-level data to predict exact flood depths for any given rainfall amount. What sets our project apart from traditional flood zone identification tools is our integration of GroundSource, a newly published dataset that incorporates historical flood events. By analyzing past flooding patterns, our model can account for real-world factors that traditional models may overlook, such as poor drainage systems and areas prone to water accumulation.

How we built it

We built FloodSight using a Python backend that integrates real-world geographic and environmental datasets, including USGS elevation data, land cover information, soil characteristics, and historical flooding records from GroundSource. Our flood modeling algorithm analyzes terrain and simulates how rainfall runoff moves and accumulates across an area to estimate potential flooding. We then use an HTTP API to connect these calculations to our application, allowing users to visualize flood risk and estimated water depth at specific locations. For the frontend, we used SwiftUI to create an interactive iOS application, integrating MapKit to display flood zones and identify areas at risk. We also incorporated RealityKit to create an augmented reality experience that allows users to visualize rising water levels in their surroundings.

Challenges we ran into

One of our biggest challenges was integrating environmental datasets with different formats, resolutions, and geographic coverage. We also struggled with accurately modeling coastal flooding, where rising sea levels introduce complexities beyond rainfall alone. Another major challenge was debugging our flood calculations, as some locations returned zero flood depth despite extreme rainfall.

Built With

• arkit
• coreml
• docker
• figma
• groundsource
• http
• mapkit
• openstreetmap
• python
• realitykit
• ssurgo
• swift
• swiftui
• usgs
• vdatum