Inspiration
Navigating through the classes of a big academic institution, such as Johns Hopkins can be hard, especially for someone studying remotely, or who doesn't have a distinct direction of studies. We wanted to write some code to help students with their academic decisions by making the endless list of courses more accessible.
What it does
ClassMatch can find similar classes to the ones you have already taken and suggest them to you. To do that, it sources for class descriptions online, and compares them with all others. The user can provide feedback to these suggestions which is taken account for future improvement. For now, we are focused on classes offered by the WSE JHU school.
How we built it
We built the text processing algorithm, keyword finding and similarity metric using Python (main packages numpy, scipy, sklearn, nltk) and the html/css interface using flask. Our data were sourced from the JHU SIS API.
Challenges we ran into
Extracting the data from the Johns Hopkins SIS API was tricky, as well as using Flask to connect html with python.
Accomplishments that we're proud ofclass hacking
ClassMatch has basic functionality and we're proud that it works! It provides class suggestions that make sense, independent of department. Extracting all data from the JHU SIS API effectively is an accomplishment by itself.
What we learned
This has been an educational journey for us. We learned a lot about NLP, data structures, html and of course, Johns Hopkins classes itself.
What's next for ClassMatch
We are working on adding further functionalities: 1) Searching classes based on interests 2) Matching classes from different schools (transfer students, international students) 3) Extending ClassMatch for research opportunities (LabMatch) 4) Extending ClassMatch for job opportunities (JobMatch)
Log in or sign up for Devpost to join the conversation.