AI Sports Odds Web Application

Goal

Build a clean, fast web app that:

• Generates AI predictions and “fair prices” for every game.
• Scrapes and compares live bookmaker odds in one screen.
• Tracks price movement and flags value bets.
• Lets users log bets and see ROI/performance.
• Incorporates weather into models.

Sports & Markets

• Leagues: NRL, AFL, NBA, NFL, NHL, MLB, NBL
• Markets: H2H win probability, Total (points), Line (spread)

Core Features

1. AI Predictions
• H2H win probabilities (Random Forest Classifier).
• Total & Line predictions (Gradient Boosting Regressor).
• Inputs include team strength, form, home/away, weather, historical stats.
• Outputs “fair odds” per market; models update on new results.

2. Live Odds Comparison (Scraped)
• Bookmakers: Sportsbet, TAB.com.au, Ladbrokes, Tabtouch, Betfair, BetRight, Picklebet (extensible).
• Display all prices per market; auto-highlight best price.
• Data refresh multiple times daily (cron/queue).

3. Price Movement History

• Store opening price and every change with timestamps.
• Show movement table to identify steam.

4. Value Bet Suggestions
• Compare fair price vs. best available; flag if edge ≥ threshold (default 15%, user adjustable).
• “Suggested Bet” badge/alert.

5. Bet Tracker
• Manual entry: selection, market, odds, stake, bookmaker, result.
• Auto P/L, ROI, exposure, per-bet summary.

6. Weather Integration
• Fetch forecast by game location and feed into prediction pipeline.

Web App Requirements

• Modern, responsive UI with team logos, sport filters, odds toggles (H2H / Total / Line), and a panel for Suggested Bets & Price Moves.
• Basic auth and roles (User/Admin) acceptable; simple onboarding.
• Server jobs for scraping/model updates; retry & rate-limit handling.
• Persistent storage for odds snapshots, bets, and results.

Tech (suggested, flexible)

• Frontend: React/Next.js
• Backend: Node.js or Python (FastAPI/Express) + job runner (Celery/BullMQ)
• ML: scikit-learn (RF/GBR) with versioned models
• DB: Postgres (primary)
• Weather: standard forecast API integration

Deliverables

• Deployed web app (cloud) with env-based config.
• Source code (repo) + README, setup scripts, and seed data.
• Scrapers for listed bookies, model training scripts, and scheduled refresh.
• Minimal admin view for monitoring jobs/feeds.
• Compliance notes (see below).

Please demonstrate your capabilities with a small MVP; ongoing work will follow based on results.