🏀 NBA Sharp Analytics Platform

Professional NBA betting analytics with fully automated daily updates

Features

✅ Fully Automated - Zero maintenance required ✅ Daily Updates - Fresh predictions every morning at 9am ET ✅ Performance Tracking - Real win/loss record (no fake stats) ✅ 6 Prediction Models - Sully Four Factors, ELO, Net Rating, Pythagorean, Ensemble ✅ +EV Calculator - Expected value for every bet ✅ Multi-Book Line Shopping - Compare 5+ sportsbooks ✅ Sharp Money Indicators - RLM, Steam Moves, CLV tracking ✅ 100% Free - No subscriptions, no paid APIs

How It Works

Automated Daily Flow:

9:00 AM ET - GitHub Actions runs:

1. Fetches today's NBA schedule
2. Gets injury reports from ESPN
3. Pulls betting lines from multiple books
4. Calculates predictions using 6 models
5. Stores data in data/ folder

2:00 AM ET (after games finish):

1. Fetches final scores
2. Checks predictions vs actual results
3. Updates win/loss record
4. Calculates accuracy stats

When You Visit: Fresh data from latest automated run!

Quick Start

1. Visit: https://lukegranto23.github.io/nba-sharp-analytics/
2. See today's predictions automatically
3. Check back tomorrow for new games

Tech Stack

• Frontend: HTML/CSS/JavaScript
• Backend: Python + GitHub Actions (cron jobs)
• Database: JSON files in data/ folder
• Hosting: GitHub Pages (free)
• APIs: balldontlie.io (free NBA data)

File Structure

.
├── index.html              # Main app
├── data/
│   ├── todays_games.json   # Updated daily at 9am
│   ├── predictions.json    # Today's predictions
│   ├── predictions_history.json  # All past predictions
│   ├── performance.json    # Win/loss stats
│   ├── team_stats.json     # Current season stats
│   └── injuries.json       # Latest injury reports
├── scripts/
│   ├── fetch_nba_data.py
│   ├── generate_predictions.py
│   ├── check_results.py
│   └── update_performance.py
└── .github/workflows/
    ├── daily-update.yml
    └── check-results.yml

Performance Tracking

Check data/performance.json for live stats:

• Overall accuracy %
• Win/loss record
• By confidence level
• By model type

Contributing

This is a personal project, but feel free to fork and customize!

License

MIT