Inspiration
Deciding where to go to college is tough. When you're a student it seems like there is an endless list of universities to choose from. As students ourselves, we know what it feels like. Our goal is to provide easy, accessible information to students all around the world. With College Collage, discovering where you belong is easy.
What it does
College Collage allows users to enter their preferences regarding various aspects of colleges. Then, it compares their preferences with over a thousand colleges and returns the top 5 colleges that match their preferences. It displays the colleges with relevant information such as location and tuition in an aesthetically pleasing collage.
How we built it
We used the Python Flask framework for the backend, where we receive and parse user input. Additionally, we clean and transform a dataset that contains over 1000 colleges and certain statistics. Then, we used cosine similarity to compare user input with those colleges and score them based on how similar they are. Finally, we ranked the colleges based on fit and sent the information back to Flask to display on the website.
Challenges we ran into
One of the biggest challenges we had was finding data. Initially, we wanted to find a dataset that contained a significant amount of students and their respective colleges so that we could train a supervised learning classifier to predict the best college for a specific user input. After we realized that this data was not readily made available, we decided to use a different dataset that contained general information about colleges.
Accomplishments that we're proud of
We are proud of our final product, and we hope to continue improving and building it in the future. We came into this hackathon wanting to make something that is impactful and functional. It was a lot of hard work, no sleep, and lots of debugging. However, we persevered and we are happy with our finished product.
What we learned
We learned how to use Flask on the backend. Additionally, we gained more experience in HTML/CSS/JavaScript and machine learning.
What's next for College Collage
We hope to improve the website by adding more preferences, developing new ways to find relevant data, and incorporate new features to the recommendation system.
Log in or sign up for Devpost to join the conversation.