← All work
→
02 — Student Performance Analytics — 2025
EduPredict

Captured live from frontend-virid-pi-53.vercel.app
Overview
EduPredict is a production-oriented student performance analytics platform. A Django REST API with pagination and throttling feeds a Next.js dashboard, while a scikit-learn pipeline retrains automatically on create, update, delete, and bulk upload. Predicted rows are tracked separately from manual marks to avoid label drift, and repeated students gain historical context features like prior marks and score deltas.
Highlights
- Automatic model retraining on create, update, delete, and bulk upload
- Training-data policy that excludes predicted rows to prevent model drift
- Repeated-student historical learning: prior marks, score deltas, and gaps as features
- CSV/XLSX bulk ingestion with validation and row limits, plus a health endpoint
Stack
- Django
- DRF
- Scikit-learn
- pandas
- Next.js 14
- TypeScript
Next project
SMS Spam Detection