🎓 AI Learning Analytics

A hackathon prototype that combines Databricks, Delta Lake, PySpark, and an LLM (Mistral / LLaMA via Databricks Model Serving) to build an intelligent learning analytics platform.

Features

Feature	Description
📊 Dashboard	Student progress overview with score trends
📝 Quiz Generator	AI-generated MCQ quizzes on any topic
📄 Summarizer	Condense notes and chapters into bullet points
📤 Data Upload	Ingest CSVs/JSONs into Delta Lake

Project Structure

databricks-hackathon/
├── config.py                  # Centralized settings (reads .env)
├── requirements.txt           # Python dependencies
├── .env.example               # Environment variable template
│
├── frontend/                  # Streamlit UI
│   ├── app.py                 # Entry point & sidebar navigation
│   └── pages/
│       ├── dashboard.py       # Analytics overview
│       ├── quiz.py            # AI quiz interface
│       ├── summary.py         # Text summarizer
│       └── upload.py          # File upload & preview
│
├── backend/                   # AI / business logic
│   ├── llm_client.py          # Databricks Model Serving wrapper
│   ├── quiz_engine.py         # Quiz generation & parsing
│   ├── summary_engine.py      # Text summarization
│   └── insights_engine.py     # Learning recommendations
│
└── data_processing/           # Spark + Delta Lake
    ├── schema.py              # PySpark StructType schemas
    ├── pipeline.py            # Data ingestion to Delta
    └── queries.py             # Reusable analytics queries

Quick Start

1. Clone & install

cd databricks-hackathon
cp .env.example .env           # ← fill in your Databricks credentials
pip install -r requirements.txt

2. Run the Streamlit app

streamlit run frontend/app.py

3. Connect to Databricks

Set these values in your .env:

Variable	Description
DATABRICKS_HOST	Your workspace URL
DATABRICKS_TOKEN	Personal access token
DATABRICKS_WAREHOUSE_ID	SQL warehouse ID
LLM_ENDPOINT_NAME	Model Serving endpoint name
CATALOG_NAME	Unity Catalog name
SCHEMA_NAME	Schema for learning tables

Tech Stack

• Frontend: Streamlit
• Backend: Python + OpenAI-compatible client
• Data: PySpark + Delta Lake on Databricks
• AI: Mistral / LLaMA via Databricks Model Serving
• Catalog: Unity Catalog