LoamGnome Garden

Plant monitor registration
Physical prototype with sensor array
personalized AI recommendation that responds to real-time sensor data
Welcome splash

Inspiration

The greenery decor of the stage and presentations inspired us to lean into an organic, earthy theme for our project. Seeing the options of hardware sensors available to us, we wanted to combine them all in one project.

What it does

LoamGnome Garden is an AI-powered app that monitors the health of your garden and gives you natural, tailored recommendations powered by Google Gemini.

How we built it

For our hardware element, we used an Arduino UNO R4 WiFi, for its IoT capabilities. We combined a variety of sensors from the hardware lab. Our backend is written in Django, which handles storing data and taking updates from the hardware edge device. Our frontend is a React app hosted on Vite that handles web requests with ElysiaJS. Our data is handled by TigerData, which made setting up our database a breeze.

Challenges we ran into

We struggled a lot with learning backend development and building the networked element of the hardware sensors. This was our first time building IoT hardware, and we had a lot of gaps to cover in processing inbound data.

Accomplishments that we're proud of

We're proud of how we were able to divide up and parallelize the workload for the project, which is something each member was either inexperienced with or had struggled with at previous hackathons. This was our first app built with microservices and IoT integration.

What we learned

We learned a lot about database integration with Django. It was a real-world experience of integrating the frontend, using a backend for the frontend, with the backend server. Using a microservices model was new to us, and we found how much it simplified development when integrated well. Additionally, this was our first IoT hardware project, so it was fun learning how to implement a HTTP messaging system with our C++ Arduino sketch.

What's next for LoamGnome Garden

Develop bespoke hardware that elegantly fits in with plants, expand sensor collection, and augment how we feed Gemini data to get better, actionable natural-language recommendations.