Inspiration
I learned about walkable cities and neighborhoods through the internet with channels like CityNerd and NotJustBikes. My peers and I grew frustrated with how hostile Florida was to pedestrians and wanted to create a way to quickly suggest improvements for future development.
What it does
This app takes in two addresses from OpenStreetMaps to form a route. This route is then broken up into 500m segments, with a picture per segment pulled from Mapillary. These images are then fed into our AI agent, Jarvis, which analyzes them for any missing features (greenery, sidewalks, bike lanes). The before and after is then displayed on the website, alongside the suggestions.
How we built it
This was built with HTML/CSS, JavaScript, Java SpringBoot, Ollama, and Figma.
Challenges we ran into
The AI was surprisingly slow, for some routes it would take several minutes to analyze. OpenStreetMaps was also annoying to work with. It would stall if too many people were making a request and some of the routing data was very outdated. Both issues were solved by caching the routes when they did work.
Accomplishments that we're proud of
Ashley: I'm proud of myself for sticking with the mentor position. I feel like I actually made an impact. Tiffany: I think improving on JavaScript for me!! This project helped me a ton in becoming more familiar with the language since I would consider myself really beginner with it at the start :O Mariana: I really loved to see our work being put together !! Grace: Making a leaflet map for the first time was fun.
What we learned
Ashley: I finally learned how to use restful APIs. I've never done anything backend before. Tiffany: I learned a lot more about back-end development!! Even though I wasn't directly working on it, I had absolutely no idea what back-end entailed before...and APIs... Mariana: I also learned a lot about creating an AI agent and how they are used for different reasons Grace: I got to learn more about javascript and frontend development
What's next for GreenRoute
I hope to keep working at it to make it faster and give it more things to analyze. Hopefully one day it will become an app with real users.
Built With
- css
- figma
- html
- java
- javascript
- ollama
- springboot