Neighborhood Sprout

Inspiration

In a time when screens often replace scenery, we forget to simply touch grass. Our phones connect us to the world—but what if they could help us reconnect with the one right outside our door? This nature photo walk encourages participants to explore their neighborhoods, notice the unnoticed, and celebrate the role of nature in shaping community and belonging.

What it does

Neighborhood Sprout is an interactive community engagement app that uses plant documentation to strengthen connections between residents while contributing to the neighborhood’s environmental data.

Through the app, users can share posts about plant discoveries—documenting not only information about the plants but also personal memories or stories connected to them. This fosters a sense of community rooted in shared experiences and local nature.

Neighborhood Sprout also features scavenger hunt–style quests that encourage users to explore and document plants in their area. Participants receive a list of plants to find and upload photos once they’ve located them.

For users unable to explore outdoors due to accessibility or safety concerns, the app offers a virtual mode with a 3D map of the neighborhood. In this mode, users can embark on a digital scavenger hunt—searching for plants within the virtual environment. When a plant is found, clicking on it reveals an informational pop-up, allowing users to learn and engage with their environment from wherever they are.

How we built it

We used Windsurf to create the app, three.js for 3D modeling, We used a YOLOv4-v3 model found on GitHub. We trained it using a 2017 training dataset from iNaturalist to identify species of plants.

Challenges we ran into

One of our main challenges was developing the virtual game components, as most of our team is new to coding and AI. Creating the 3D map proved particularly difficult, especially when integrating it with interactive features. We also faced obstacles in accessing and managing large datasets needed to train our AI model effectively, and were unable to download the AI learning agent due to Wifi issues.

Accomplishments that we're proud of

After struggling for a while, we managed to make a 3D model of a street using Google Colab and Visual Code Studio. We also finalized the camera and upload feature of the app, and were able to piece together most of the project.

What we learned

We learned how to use HTML and JavaScript to create and program an interactive 3D model of a street, how to use Windsurf to make an HTML-based web app, and learned about YOLO object detection AI and how to build data sets to train them. We also learned how to use GitHub as a resource to find tools we could use.

What's next for Neighborhood Sprout

We hope to include a reward system for users who complete the scavenger hunt quests, by adding a virtual plant that would grow as the user completes more quests over time. If the user goes for periods without doing quests, the plant can wilt and eventually die.