AI Advantage Sports

Sports analytics product for ML-driven picks, bankroll sizing, and live betting workflows.

Why This Repo Matters

This project shows product-oriented ML work rather than notebook-only experimentation:

• prediction workflows tied to a user-facing app
• live odds and betting-oriented UX
• Kelly-based stake sizing and decision support
• production-style frontend architecture with a real deployed surface

Features

• AI-Powered Analysis - Advanced machine learning models analyze thousands of data points in seconds
• Game Analyzer - Enter any matchup for instant AI analysis and betting recommendations
• Real-Time Odds - Track line movements and find value before the market adjusts
• High Accuracy - Models consistently outperform traditional handicapping methods
• Bankroll Protection - Smart stake sizing recommendations to protect your investment
• Kelly Criterion - Optimal bet sizing based on edge and bankroll management
• Multi-Sport Support - NBA, NFL, and MLB predictions
• Premium Tier - Advanced features with Stripe subscription integration

Highlights

• Live product experience at aiadvantagesports.com
• Supports NBA, NFL, and MLB workflows
• Includes odds, picks, leaderboard, and premium/paywall paths
• Useful as a portfolio example of turning modeling ideas into an actual product

Tech Stack

• Frontend: React, TypeScript, Tailwind CSS
• UI Components: shadcn/ui
• Build Tool: Vite
• Styling: Custom dark theme with green accents

Project Structure

src/
  components/        UI and product features
  lib/               predictions, rosters, Stripe helpers
  pages/             landing page, picks, leaderboard

Getting Started

git clone https://github.com/ianalloway/ai-advantage.git
cd ai-advantage
npm install
npm run dev

Then open http://localhost:5173.

Deployment

The site is configured for static hosting plus preview-ready serverless payments:

• netlify.toml keeps the current Netlify SPA deploy working.
• vercel.json adds SPA rewrites so Vercel previews can serve /daily-picks and /leaderboard.
• api/create-checkout-session.ts and api/checkout-session.ts let Vercel previews create and verify Stripe Checkout Sessions securely on the server.

Stripe Checkout env vars

For the server-created Stripe flow, set these env vars in your deployment target:

STRIPE_SECRET_KEY=sk_live_or_test_...
STRIPE_PREMIUM_PRICE_ID=price_...
STRIPE_ONE_TIME_PRICE_ID=price_...
PUBLIC_APP_URL=https://your-preview-or-production-domain

If those are missing, the app falls back to the static Stripe Payment Links from VITE_STRIPE_CHECKOUT_URL and VITE_STRIPE_ONE_TIME_CHECKOUT_URL.

Newsletter automation env vars

The homepage email form now posts to /api/newsletter-subscribe. On each signup it:

• creates a child page in Notion with the subscriber details
• sends you a notification email
• redirects the visitor into the official Substack subscribe page

Set these env vars before turning the form on in production:

NOTION_API_KEY=secret_your_notion_integration_token
NOTION_PARENT_PAGE_ID=your_notion_parent_page_id
RESEND_API_KEY=re_your_resend_api_key
RESEND_FROM_EMAIL="AI Advantage <onboarding@yourdomain.com>"
NOTIFY_EMAIL=ian@allowayllc.com
SUBSTACK_PUBLICATION_URL=https://allowayai.substack.com

NOTION_PARENT_PAGE_ID should be the page where you want each new subscriber page to appear. Share that Notion page with your Notion integration before testing.

Notes

• This repository is best read as a shipped product and frontend/application example.
• Modeling details live in adjacent public projects such as sports-betting-ml, nba-ratings, and nba-clv-dashboard.

Author

Ian Alloway

• Portfolio: ianalloway.xyz
• LinkedIn: linkedin.com/in/ianit
• GitHub: github.com/ianalloway

License

This project is proprietary. All rights reserved.