iGradeU: CourseGradePredictor

Inspiration

As students, we all share a common experience—worrying about our academic performance and feeling uncertain about our future. Some of us underestimate our abilities, while others overestimate them and underprepare. What unites us is the stress of not knowing where we stand until it’s too late.

We asked ourselves: how can we use the skills we’ve gained—data extraction, statistics, and AI—to make school less stressful and more predictable? That question led us to build iGradeU, an academic foresight tool that helps students turn guesswork into clarity.

What It Does

iGradeU combines a student’s Canvas performance data, a course’s syllabus, and RateMyProfessor insights to predict how well the student is likely to do in a specific class.

Predicts a final grade forecast for a chosen course
Provides student skill scores (e.g., exams, projects, participation, assignments)
Rates the course’s difficulty and teaching style using RMP data
Gives a clear verdict: what you’ll excel in, where you might struggle, and whether the course is a good fit

In short: it’s your personal academic coach, powered by real data.

How We Built It

Frontend: React + Vite
Backend: Django REST Framework
APIs:
- Canvas API for grades, weights, and assignment categories
- RateMyProfessor API (adapted from unofficial sources) for difficulty and teaching ratings
- OpenAI API for natural language processing of syllabi and predictive modeling

We engineered prompts and custom algorithms to guide OpenAI in producing statistically sound scores and personalized feedback, synthesizing raw data into clear recommendations.

Challenges We Faced

API Limitations: Canvas was new to us, and RateMyProfessor had no official API, forcing us to adapt and manipulate an outdated unofficial one.
Prompt Engineering: It took extensive trial-and-error to find prompts that produced reliable, statistically valid outputs.
OpenAI Rate Limits: Heavy usage meant hitting token caps; we had to manage credits and optimize requests.
Collaboration Complexity: With diverging GitHub branches and a time crunch, merging code and cleaning up the frontend tested our teamwork.
Frontend Polish: Styling and UX clarity were rushed, leaving room for improvement.

—

Accomplishments We’re Proud Of

Successfully modified and repaired an unofficial API to make it functional.
Designed prompts that produced reliable, data-driven predictions.
Built a working grade prediction system end-to-end in a short time frame.
Turned an ambitious idea into a functioning prototype that directly addresses a real student problem.

What We Learned

Teamwork is everything: Effective communication and role delegation made us far more productive than working solo.
Adaptability matters: From API limitations to token caps, solving problems on the fly was both challenging and rewarding.
We can create impact: This project proved to us that our skills can be used to tackle the same struggles we face as students every day.

What’s Next for iGradeU

This is just the beginning. Future directions include:

Improved UX/UI: A clearer and more polished user experience with intuitive design.
Deeper RMP Integration: Expanding beyond basic ratings to include sentiment analysis of reviews.
Stronger Predictive Models: Incorporating more advanced algorithms and larger data sets for higher accuracy.
Scalability: Making iGradeU usable for students at multiple universities, not just our own.