Product blueprint
This implementation uses GenMB Authentication and user-scoped KV persistence for the interactive workspace. A production institution deployment should move shared class records and authorization checks to server-side functions backed by a relational database with row-level policies.
Frontend: React, TypeScript, Vite, Tailwind CSS, accessible SVG charts.
Backend: serverless functions for protected mutations and score calculation.
Database: PostgreSQL with row-level security and audit fields.
Auth & hosting: GenMB Authentication and managed static/serverless hosting.
Faculty should only query classes they own or administer. Students should only query records where student.auth_user_id = current_user_id. Invite acceptance links a verified identity to a student record; it never grants roster access.
Data model
| Table | Key fields | Purpose |
|---|---|---|
| faculty | id, auth_user_id, name, email | Core class and identity relationship. |
| departments | id, faculty_id, name | Core class and identity relationship. |
| classes | id, department_id, name, academic_year, section | Core class and identity relationship. |
| students | id, auth_user_id, class_id, name, email, roll_no | Core class and identity relationship. |
| attributes | id, class_id, position 1–6, name, data_source | Core class and identity relationship. |
| daily_logs | id, student_id, topic, category, confidence, notes, created_at | Student learning activity feed. |
| attribute_scores | id, student_id, attribute_id, score 0–100, calculated_at | Fast chart reads and history. |
Protected API surface
A daily log is stored against the authenticated student, then a server-side calculator updates the relevant attribute score according to the faculty-selected source: learning log confidence, attendance percentage, project evidence, or a faculty rating. Class radar values are aggregated from those protected student scores.