Peopler

Helping you find your community one person / event at a time

Comment

Prompt - Mental Health, Community Building and Forming Connections

The prompt we wrote our solution towards was Mental Health, where we believe having a strong community and support structure is fundamental to one’s mental health. This also ties strongly with the overall hackathon theme of resilience. Most communities come together for specific events as they serve as catalysts for people to gather. We wanted to build something to help people find their community and get into them easier by matching users to people with similar interests and an event they attend. Once talking, the user can learn about the event and have someone they know to ease the anxiety of attending an event with a large group of new people.

Inspiration

The three of us hold leadership positions in our SASE chapter, our LGBT club, and our computer science clubs and we truly believe in fostering community as these have been important structures for ourselves. We've seen new members come to meetings at some of these clubs and struggle to interact with others because many people already knew each other.

What it does

The website/app matchmakes people based off their interests so they can get involved in their community and learn about events. Once matched with another user attending an event, the user can ask questions about the event and the community. To further stir conversation, we use AI to generate personalized introductions about a match and suggest topics/events to talk about. The hope is that these two people will make a connection and the new member to the event will already know someone to feel more welcome in the community. We heavily emphasized making the tool as easy to use as possible, from incorporating AI in parts where it makes sense (see below) to asking as little information as possible from our users. For example, our project is completely passwordless; users provide an email and verify ownership of that email, and that completes the authentication process.

How we built it

We used VueJS to build our frontend, Python and FastAPI as our backend API, MongoDB as our database, and OpenAI's API to leverage ChatGPT to generate intros and topics.

Incorporation of AI

For this project, we opted to not use AI tools to write code for our website and/or backend. However, we believe that AI opens up lots of new opportunities that weren't previously available and so we've incorporated AI in a few scenarios where we think it'll elevate this project.

Profile Descriptions

Our project never asks the user to write a bio for themselves, as we believe that writing one can be stressful for new users just trying to find some community. Instead, we collect a list of hobbies and interests each user has and determine matches based on those instead. We then use a large language model (LLM), in particular, GPT-3.5, to generate profile descriptions for potential matches tailored specifically towards hobbies and interests shared by both sides. We hope that this allows connections to feel more personal, and to allow one to find common ground more easily as shared interests often lead to stronger friendships.

Chat Introductions

It can be similarly daunting to reach out with a cold message to someone you're not familiar with. Using a similar concept and tooling, we've used AI to provide users with some ideas on what they can start a conversation with. These ideas are once again tailored to both users' hobbies and interests for maximum compatibility.

Challenges we ran into

We did not have the full time to work on the project as we had other commitments, so we spent only around 6-7 hours total. We were also not able to fully integrate our back-end and front-end for our project in time, so we prioritized design and user flow over making our prototype 100% accurate.

Built With

Share this project:

Updates