CoursifyUMD

Have you ever felt confused about what courses you should take? CoursifyUMD is an app that allows users to select details about themselves and generate a list of courses that suit them the most

Comment

Inspiration

Students have a difficult time determining what courses to take. We wanted to create a solution that could recommend courses most tailored to them.

What it does

Our app allows users to select options about themselves regarding their major, year, personal interests, and desired industry field, then recommends several lists of courses to take that fit them. These lists are divided by the individual characteristics entered by the user. There is a list of courses based on their major, a list based on their personal interests, and a list based on their desired industry field.

How we built it

We used a Python Flask application to organize the backend and React to create the front end. We created a script to web scrape data from the Testudo and PlanetTerp websites, embed the data using OpenAI embeddings, and stored the resulting vectors in a Pinecone vector database. We also created a Supabase database to store user data and their queries.

Challenges we ran into

Because we worked on the Backend and Frontend separately, we had a difficult time integrating them together. We were ultimately able to overcome this problem by discussing each section thoroughly and properly explaining how data is sent across platforms.

Accomplishments that we're proud of

We are proud to create an application using a React frontend and Flask backend, creating a secure database to house user data, and constructing a vector database to query results.

What we learned

We learned to work with technologies such as OpenAI embeddings, vector databases, database querying, and general frontend and backend development.

What's next for CoursifyUMD

In the future, we would like to implement more personalized features for users to get more generated curated lists. We'd also like to look into other algorithms that could provide even more relevant results.

Built With

Share this project:

Updates