Inspiration

Lecture halls aren't the greatest way to meet new friends, and it shows.

Loneliness is a huge issue in universities; in 2017, 63.1% of nearly 48,000 college students indicated that they had felt “very lonely” in the previous 12 months [1]. College students often struggle to make friends outside of dorms or clubs, and while juggling all the pressure of classes and extracurriculars, it can be hard find friends - particularly because there isn't a great way to do so. At the same time, studies show that people tend to underestimate the mental benefits brought by conversing with strangers [2].

Fifteen aims to leverage the unexplored potential of with-stranger conversations to improve the well-being of students and reduce loneliness on college campuses. We aspire to bring college students closer in an enjoyable, adventurous way and cultivate a tighter-knit student community.

Fifteen won't just be for college students, either. Fifteen can be a tool to help build community everywhere. You might have heard of the six degrees of separation; everyone in the world is connected through no more than six connections. Through Fifteen, we hope to bring this number down to five.

What it does

Fifteen is a platform that connects university students and encourages them to meet up with another random student for fifteen minutes every day. We use our algorithm to pair up our users to a stranger they'll get along with, and then the magic starts. They first chat with each other online anonymously, in a low-pressure, low-commitment environment, and then decide if they want to spend fifteen minutes together in real life.

How we built it

  • Dynamic website built off HTML, CSS, JavaScript, Python, Flask, Node.js, and Socket.io
  • Custom design (all images are hand drawn and color-blind friendly)
  • Functional chat feature built with NodeJS Express framework embedded into the website
  • NetworkX and Geopandas backend processing for matching algorithm
  • Google Firebase database for user profiles
  • Google BigQuery for GIS data along with Cloud GPUs for processing
  • Matplotlib and imagemagick for creating visuals
  • Researched Meyer-Briggs personality tests to create entry survey

Challenges we ran into

  • Dream bigger — we started with a big idea to address the college student mental health issue, spending a lot of time doing user research and brainstorming the most effective and realistic solution.
  • Algorithm & Visualization/database —figure out how to diagram and deliver this to our user, computation times were resource consuming and difficult as the dataset got larger
  • Web development — Integrating multiple languages together, specifically with Flask and NodeJS; Making our website response and mobile-friendly
  • User survey — challenging to balance between the comprehensiveness, length, and style during user survey design.
  • Design — a lot of factors are put into considerations to increase Fifteen’s accessibility, user-friendliness, and desire to use our product
  • repl.it — a coding collaboration platform that is always broken.

Accomplishments that we're proud of

*Human-centered design process - our product design iterates through cycles supported by substantial user research, which includes but is not limited to an anonymous chat function, color-blind friendly color palette, optimization of user flow, fifteen-minute meetup time and visualization of data *Interdisciplinary project —with diverse backgrounds, we capitalize on our engineering skill, design knowledge, and social science research methodologies to build Fifteen

What's next for Fifteen

*Expanding target audience - the elderly, middle-aged workers, etc. More functions - give users the ability to choose between exploration mode and similarity mode (more flexibility in who they match with) *More incentives - find sponsorships to provide things like free meals to incentivize users to meet offline Chatbot AI - suggest conversation topics and fun things to do to help reduce planning anxiety

References [1] https://www.acha.org/documents/ncha/NCHA-II_FALL_2017_REFERENCE_GROUP_EXECUTIVE_SUMMARY.pdf [2] https://psycnet.apa.org/record/2014-28833-001

Share this project:
×

Updates