Video Inspiration Finder 🎯

AI-powered YouTube video recommendation system that learns your preferences to suggest coding videos you'll love.

Features

🔍 Smart Search: Finds viral coding videos from YouTube
🤖 Machine Learning: RandomForest model learns your preferences
📊 Web Dashboard: YouTube-like interface showing personalized recommendations
🔒 Privacy First: All data stored locally in SQLite
⚡ Fast Setup: One command installation and execution

Quick Start

Option 1: Command Line Interface

./setup.sh

This will:

Create virtual environment
Install dependencies
Search for videos
Start interactive rating session

Option 2: Web Dashboard

./setup.sh  # First time setup
python run_dashboard.py

Opens a YouTube-like dashboard at http://localhost:5000

How It Works

Search Phase: Queries YouTube API for trending coding videos
Rating Phase: You rate videos as like/dislike with optional notes
Learning Phase: After 10+ ratings, ML model activates
Recommendation Phase: AI suggests videos based on your preferences

Project Structure

src/
├── database/     # SQLite operations
├── youtube/      # YouTube API integration  
├── ml/          # Machine learning pipeline
└── rating/      # Interactive rating system

main.py           # CLI application
dashboard_api.py  # Web API server
templates/        # Dashboard HTML/CSS/JS

ML Pipeline

Features: 11 extracted features (keywords, engagement, sentiment)
Model: RandomForest with 100 trees
Training: Incremental learning after each rating
Prediction: Confidence scores for video recommendations

Requirements

Python 3.7+
YouTube Data API v3 key
SQLite (included with Python)

Setup

Get YouTube API key from Google Cloud Console
Create .env file:
```
YOUTUBE_API_KEY=your_api_key_here
```
Run ./setup.sh

Dashboard Features

📱 Responsive YouTube-like design
🎯 AI confidence scores for each recommendation
🔄 Real-time model status (learning vs trained)
🖱️ Click videos to open in YouTube
📊 Visual feedback on model training progress

Commands

./setup.sh              # Full setup and CLI
python main.py           # CLI only
python run_dashboard.py  # Web dashboard
python dashboard_api.py  # API server only

Built with Python, Flask, scikit-learn, and YouTube Data API v3.

2.5 KiB Raw Blame History