Academic Guidance & Monitoring Intelligence System
An intelligent academic data platform engineered with FastAPI, PostgreSQL, and scikit-learn to score student performance risk and deliver automated guidance.
Why this matters
Educational institutions collect vast amounts of academic data—grades, attendance, and assessment scores—which remain siloed in static spreadsheets. Risk factors for student probation or dropout are often discovered too late for effective intervention.
Understanding the Context
Interviews with faculty and institutional leadership revealed that students, teachers, and administrators view isolated fragments of performance data. Creating a unified intelligence layer bridges this gap, giving stakeholders proactive early warnings.
Different perspectives, one system
Student Dashboard
Personalized analytics showing current trajectory, grade momentum, risk indicators, and customized study guidance.
Faculty View
Class-wide monitoring dashboard identifying struggling students early to enable targeted academic support.
Executive Admin Dashboard
Institution-wide reporting on overall performance trends, course pass rates, and policy impact.
System Design
Lightweight FastAPI backend connected to a PostgreSQL database. Data pipelines normalize raw ERP records, extract rolling performance features, and feed a scikit-learn model to output risk scores and rule-based advice.
Handling data accurately
- Automated data ingestion from standardized CSV/ERP exports into PostgreSQL.
- Data cleaning, missing value handling, and grade scale normalization.
- Feature engineering computing rolling grade momentum and attendance risk factors.
Machine Learning & Agents
- Supervised machine learning classification for student risk scoring.
- Rule-based expert decision system generating actionable guidance recommendations.
Dashboards & Interfaces
Student View
Performance metrics, risk status, and personalized study milestones.
Faculty View
Class roster heatmaps and early warning alerts.
System in action
Future Improvements
- Integration with Canvas and Moodle LMS webhooks for real-time grade sync.
- Deep learning sequential models for long-term degree completion forecasting.
- Automated notification engine pushing risk alerts via email and WhatsApp API.
Ready to dive deeper?
Explore the architecture, commits, and implementation details of this project.