Professor's AI Analytics Dashboard

An AI-powered backend service designed to provide university professors with deep, actionable insights into their lectures.

This system:

• Transcribes lecture audio
• Uses Large Language Models (LLMs) to analyze and structure the content
• Cross-references that content against the official course syllabus to track progress and coverage
• Exposes a clean, fast API to power a frontend dashboard.

---

Core Features

🎧 Async Lecture Processing

Upload lecture audio/video files and get an immediate response. All heavy processing (transcription, AI analysis) happens in the background.

🧠 AI-Powered Transcription

Uses OpenAI’s Whisper-large-v3 for highly accurate lecture transcriptions.

Dual AI Analysis

• Dashboard Analytics: LLM generates structured JSON for key metrics: topics covered, key points, questions asked, examples used, and a high-level summary.

• Pedagogical Notes: A second, detailed LLM call generates rich, publication-quality class notes in JSON.

📘 AI Syllabus Parsing

Upload a course syllabus (PDF or DOCX) — Llama 3.3 reads and parses it into a day-by-day JSON “course roadmap”.

📊 Syllabus Coverage Tracking

Automatically compares actual topics from lectures vs planned topics from the syllabus. Outputs a detailed coverage report (e.g., 90% covered, 2 topics missing).

Analytics API

Full suite of API endpoints to power the frontend dashboard: /analytics/dashboard, /analytics/questions, and more.

🔍 NLP Topic Modeling

Includes Gensim/NLTK to perform LDA topic modeling on transcripts as an alternative analysis layer.

---

Tech Stack

Layer	Technologies
Backend	FastAPI, Uvicorn
Database	SQLAlchemy, SQLite
Validation	Pydantic
AI & ML	OpenAI (Whisper, Llama 3.3), Gensim, NLTK
File Parsing	PyPDF2, python-docx
Async	BackgroundTasks (FastAPI)

---

⚙️ Setup & Installation

# 1. Clone the repository
git clone https://github.com/sibi-seeni/professors-dash.git
cd professors-dash

# 2. Create and activate a virtual environment
python -m venv .venv
source .venv/bin/activate
# On Windows:
# .venv\Scripts\activate

# 3. Install dependencies
pip install -r requirements.txt

🔑 Create Your Environment File

Create a file named .env in the root directory and add your keys:

OPENAI_API_KEY="sk-..."
OPENAI_API_BASE="https_your_proxy_url_com"

▶️ Run the Application

uvicorn main:app --reload

Your API will be live at: 👉 http://127.0.0.1:8000

Interactive Docs: 👉 http://127.0.0.1:8000/docs

---

🧩 Key API Endpoints

Method	Endpoint	Description
POST	/upload/	Upload a lecture audio/video file (triggers background processing)
POST	/upload_syllabus/	Upload a syllabus PDF/DOCX (triggers AI parsing & coverage)
GET	/lecture/{lecture_id}	Retrieve lecture data (transcript, analytics, notes, etc.)
GET	/lecture/{lecture_id}/notes	Get clean JSON of pedagogical class notes
GET	/analytics/dashboard	Get combined analytics for dashboard UI
GET	/syllabus_result/	Get latest syllabus coverage report with AI-generated roadmap

---

🪄 Next Steps: The Road to Canvas Automation

The current workflow requires manual uploads. The next phase integrates with the Canvas API to make the process fully automated — turning this into a hands-free Teaching Assistant.

1. 🔐 Authentication

• Add canvasapi to requirements.txt

• Add to .env:

  CANVAS_API_URL="https://ufl.instructure.com"
  CANVAS_API_KEY="your_key_here"

• Create a helper function to initialize the Canvas object.

---

2. ⬇️ Pull from Canvas (Automated Ingestion)

Replace POST /upload_syllabus/

New endpoint:

POST /sync_syllabus/{course_id}

• Uses canvas.get_course(course_id) to locate the syllabus file.
• Downloads it and passes it to syllabus_tracker.process_syllabus_file.

Replace POST /upload/

New endpoint:

POST /sync_lectures/{course_id}

• Lists files in a Canvas folder (e.g., “Lecture Recordings”).
• Detects new files and queues them for background processing with processing.process_lecture_file.

---

3. ⬆️ Push to Canvas (Automated Publishing)

After background tasks complete:

📝 Auto-Publish Class Notes

• Format notes_json into HTML.
• Use course.create_page() to publish new “Lecture Notes” pages for students.

❓ Auto-Create Draft Quizzes

• Convert quiz_json into QuizQuestion format.
• Use course.create_quiz() to build new unpublished quizzes for professors to review.

📈 Update Professor’s Dashboard

• Use syllabus_tracker’s coverage data.
• Create or update a private “Instructor Dashboard” page (published=False) inside Canvas.

---

🧭 Vision

By integrating with Canvas, this project evolves from a manual dashboard to an autonomous analytics companion for educators — tracking lecture content, syllabus progress, and learning outcomes seamlessly.

---