Recycle Master

Inspiration

Amidst the Covid-19 pandemic, the world has centralized to live a hybrid focused lifestyle. It is more important now than ever for humans to understand where to dispose their waste. Improper waste disposal contaminates the water, harms the local ecosystem, and increases greenhouse gas emissions. Disposing your waste correctly is an important factor in helping the world reach net zero emissions and ultimately saving the planet.

What it does

Recycle Master is a tool created for ToHacks22 where you are able to ask our friendly robot the Recycle Master whether your object belongs in the Blue Bin, Green Bin, Garbage Bin, Household Hazardous Waste, Drop Off Depot, Electronics, Metals, Oversized Item, Yard Waste, or is Not accepted.

How we built it

This project was built using next.js, and NLP from co:here to creatively respond to various disposable items. We utilized data from the City of Toronto to train the classifier model from co:here to generate responses and display the data through an interactive chat window. We implemented authentication to create a more personal experience with the user. Our friendly robot Recycle Master was created using Photoshop and animated using TailwindCSS.

Challenges we ran into

While finetuning our model using co:here, we had to extract and convert all the data from the City of Toronto waste management website to a csv file. This was a challenge because manually entering the data was not feasible. To overcome this challenge, we used TypeScript to convert all the data from the website into an acceptable csv format for co:here to finetune.

Accomplishments that we're proud of

Our biggest accomplishment is learning, understanding and implementing the NLP finetuned model from co:here to accurately classify where various items should be disposed. We were also proud of creating API requests and connecting the back end with the front end of the project.

What we learned

During this project we collectively learned to use co:here and NLP to create an A.I bot. We also learned form submissions and passing states through parameters. We also learned to creatively implement the front end and back end aspects of the project.

What's next for Recycle Master

Next for Recycle Master is to use the sign in feature to store past user data so the individual can check their item history. Furthermore, we want to implement an image processing system that correctly categorizes the item to its bin based on its features.

Vincent Fong, Dhruv Sharma, Christopher Fong