Inspiration
A close friend of ours struggled through his Computer Science course and ultimately failed due to a lack of timely academic support. He was bright and hardworking, but without early intervention, he fell behind and couldn't recover in time. This experience made us realize the importance of proactive academic tracking. Our goal is to prevent similar situations by providing early warnings and intervention strategies for at-risk students.
What it does
It's like an academic advisor that tracks a student's grades in real-time, and using a ML model it gives us the probability of the student failing the course and if the student is close to failing / reaching that threshold the website aims to alert the student and their advisor for early intervention.
How we built it
We used React and CSS for frontend, Python for backend, we trained our own ML model which was based on logistic regression and trained on synthetic data we generated. We also used OpenAi to parse the syllabus and give output to continue our calculation on failure value by being adaptive to the syllabus PDF.
Challenges we ran into
We originally wanted to use a quantum annealer to group clusters of students based on their failure probability score into tutoring groups, but the API key we planned on using wasn't usable anymore.
Accomplishments that we're proud of
Learning and implementing logistic regression
What we learned
We learned how to train an ML model on synthetic data we made.
What's next for ClassQGuard
We hope to implement Brightspace API for large-scale student academic advising and to take data (grades, syllabus) directly from there.
Built With
- javascript
- logisticregression
- openai
- python
- react
Log in or sign up for Devpost to join the conversation.