-
-
Interactive terrain model with risk analysis
Inspiration
- Seeing storms in NC are to this day affected and suffering by severe weather
- Friends and family impacted by the ongoing aftermath of Helene
- Communities that will never be the same
- Preventable fatalities and injuries
What it does
- Terrain map to view costs, damages, and fatalities due to previous storms and severe weather incidents
- Interactive learning experiences broken down by user base:
- Residents can educate themselves on local weather hazards and how to prepare; also includes resources to reach out to their legislators (equipping them for community advocacy)
- Local government officials and emergency managers can update themselves on recent trends and predicted issues for their county/area; also able to generate infographics to distribute to the residents under their care
- State legislators and NC General Assembly members can quickly collect facts and explore additional resources for informed decision-making
How we built it
- AI: Technical backbone to create a usable, competitive, productive worth using
- User interface
- Debugging
- Polishing code
- Humans: Producing ideas, testing concepts, writing code for beginner ideas, and bouncing ideas off one another
- Machine Learning model (gradient-boosted classified and random forest classifier) -- Vyomi S. and Ashna L.
- Exploratory analysis and visualization design -- Neha P. and Shreya D.
Challenges we ran into
- Dealing with huge longitudinal datasets, joining three separate file categories across 27 years
- Missing valuable predictors from certain datasets, so have to find supplemental datasets (cited below)
- Researching the context and use case of a topic we are not experts on to understand our audience before we create something for them
- Outputting predictive model results in a usable, accessible manner (not too technical, not too simplified)
Accomplishments that we're proud of
- Our teamwork and project planning, where everyone's ideas were heard and integrated into the project
- The fun we had during the process!
What we learned
- Generative AI is capable of a lot, but...
- Relying on it fully takes the fun out of coding and projects :)
- Its products might not hold up to enterprise- or government-grade needs and should be tested extensively
- It doesn't have as creative ideas as we do!
Log in or sign up for Devpost to join the conversation.