This was a project built at the first ever hackathon I attended (virtually tho). It was organized by BOLT club at McGill in September 2020.
Some courses in my first semester consisted of formation of groups so that students can collaborate together and have more of an in-class like experience, while being Zoom ‘24.
I feel that professors and/or TAs end up spending a lot of time in making groups if they do everything manually, or if they randomly do the group-making process, the groups usually become uneven and students don’t end up enjoying the course.
So, the solution was to kill two birds with a single arrow (SIKE, just figuratively) and make an app that does both: automates the process as well as keep the team/group chemistry as best as it can!
PathLink is an ML-powered functional prototype that automates the process of making groups based on student’s preferences, time-zones, etc. It uses the K-Means algorithm to form clusters in the data. The app itself is built using the Streamlit Python library and uses Plotly to display visualizations.
Currently, PathLink falls into the typical hackathon-project category and is just designed to work on the data type specifically collected for the project.
However, in future, I aim to make it scalable and able to fit to more generalized form datasets so that the targeted users can actually use it.
The project repo can be found here.