GreenSeat

• 

  Logo

• 

  Signup page

• 

  Matching page

Inspiration

Climate change is a ever increasing problem encroaching onto our daily lives. It is easy to push back the problem onto others with the belief that climate change can and will be solved in the future, but our group decided to work together to invoke a way to alleviate the problem, and GreenSeat is that solution. With GreenSeat we can give the world a chance for a better future.

What it does

GreenSeat is a platform that matches users to carpool together based on a variety of factors related to need and personal preference, with the majority of the weight applying to distance from one another and arrival/departure time differences. After signing up, a user is able to view basic information relating to other nearby users regarding their need for a carpool and tastes when in the car on subjects such as music, temperature, and noise level. When two users match they are able to directly message one another to work out logistics.

How we built it

We implemented a NextJS app with TypeScript to build our frontend, and used NextJS API Routes to connect with our MongoDB database hosted on DigitalOcean. We used the suite of features NextJS comes with to easily implement the authentication and routing that we needed for a multi-screen app. To add more functionality to our app, we used Googles Gemini Embedding 1 model to embed the preferences of users to a vector, and then used cosine similarity to semantically compare the similarity of each user's preferences. This similarity, alongside user distance and time differences, are all components of our ranking algorithm. Our app was hosted on DigitalOcean App Platform, which let us deploy code changes easily when we pushed to our GitHub repository.

Challenges we ran into

The biggest challenge we ran into was that some of our ideas were not feasible to implement in a 24 hour period. Scaling and designing the app in a way that it could be released to the public would take far too much time for this hackathon, but with a longer window we would be able to implement most of if not all of our future ideas.

Accomplishments that we're proud of

What we're most proud of is the way we were able to match users. We came up with an equation to optimize matches for carpooling based on time, distance, and a variety of personal preferences for a strong compatibility rating system.

What we learned

What we learned was how to leverage Artificial Intelligence to improve the speed at which we could prototype apps. AI has improved drastically over the past few years, and while manual corrections are still needed, this tool is becoming more impressive by the day.

What's next for GreenSeat

Expanding the case usage of GreenSeat (such as for school) Group Messages Better password management/software security The ability to accept/decline messages GPS integration Dynamic cutoffs for distances based on environment Profile pictures Add vector database for more efficient semantic matching Implement websockets to improve speed of messaging

Built With

• digitalocean
• gemini
• mongodb
• next.js
• react
• typescript
• vscode