Footdata - Football Match Predictions & Analytics

Overview

Footdata is an interactive football prediction platform that empowers users to build and test their own match prediction models. Rather than presenting black-box predictions, the application lets users explore how different factors (attack strength, defense, form) influence match outcomes through an intuitive, real-time interface powered by WebAssembly.

Note: This is an active work-in-progress project exploring Rust/WASM for web applications. Many features are still in development.

The Challenge

Building a performant, transparent prediction system required:

• Achieving instant client-side calculations for real-time interactivity
• Creating an intuitive model-building interface for non-technical users
• Processing and normalizing data from multiple football leagues
• Balancing model simplicity with prediction accuracy
• Delivering a native-like web experience with Rust/WASM

Technical Implementation

Frontend (Dioxus + WASM)

• Dioxus 0.6 framework for reactive UI compiled to WebAssembly
• Client-side prediction calculations running at native speed
• Interactive weight sliders for attack, defense, and form factors
• Real-time prediction updates without backend roundtrips
• Prediction breakdown showing contribution of each factor
• Zero JavaScript dependencies for core prediction logic

Backend (Axum)

• Axum 0.7 REST API for data management
• Serves team statistics and historical match data
• Backtesting calculations (backend implementation complete, UI pending)
• WIP: Automated data pipeline for league updates
• WIP: User authentication and saved custom models

Data Pipeline

• TypeScript scripts for data scraping from football APIs
• JSON data storage in shared directory structure
• Standardized schemas across leagues and seasons
• WIP: Automated real-time data scraping and updates
• Currently: Manual/periodic data updates
• Support for multiple leagues (Premier League, La Liga, etc.)

Prediction Algorithm

Multiple models implemented for comparison:

• Simple weighted scoring: Adjustable weights for attack, defense, and recent form
• Poisson distribution model: Statistical approach based on expected goals
• Monte Carlo simulation: Computational technique that uses repeated random sampling to estimate the probability of different outcomes in uncertain events
• Users can compare different prediction methodologies
• All calculations visible and explainable to users
• Scores normalized to win/draw/loss probabilities

Key Features

Currently Available

• Interactive Model Builder: Adjust prediction factors in real-time with instant feedback
• Team Comparison: Side-by-side statistics display with visual indicators
• Prediction Breakdown: Transparent explanation of why predictions are what they are
• Multiple Prediction Models: Compare simple weighted scoring vs Poisson distribution
• Multi-League Support: Data for major European football leagues

In Development (WIP)

• Home/Away Toggle: Account for home advantage in predictions
• Historical Backtesting UI: Test your model against past matches (backend ready, UI pending)
• Automated Data Updates: Real-time scraping and statistics refresh
• User Accounts: Save custom models and prediction history

Technical Achievements

• Instant predictions: Sub-millisecond calculation times thanks to WASM
• Educational value: Transparent models help users understand prediction factors
• Lightweight deployment: Efficient WASM bundle size
• Extensible architecture: Easy to add new prediction models and features
• Active learning project: Exploring Rust's ecosystem for web development

What I Learned

This project deepened my expertise in:

• Rust web development: Building full-stack applications with Rust
• WebAssembly optimization: Leveraging WASM for performance-critical calculations
• Reactive UI frameworks: Working with Dioxus's component model and signals
• API design with Axum: Building RESTful services in Rust
• Data modeling: Designing schemas for sports statistics
• Transparency in predictions: Making complex models understandable to users

Technical Highlights

Why Rust + WASM?

• Predictions calculate instantly (no server roundtrip needed)
• Memory-safe implementation prevents bugs in complex calculations
• Compiled code runs at near-native speed in the browser
• Shared business logic between frontend and backend
• Learning opportunity to master Rust's ownership system

Architecture Decisions

• Monorepo structure with shared data directory
• TypeScript for data scraping (rich ecosystem), Rust for core app
• Static JSON files over database for simplicity and portability
• Client-side calculations over backend for responsiveness

Roadmap

Short-term (In Progress)

• Complete backtesting UI integration
• Implement home/away advantage toggle
• Add automated data scraping pipeline
• Improve mobile responsiveness

Medium-term

• User accounts and saved model configurations
• Expected goals (xG) integration
• Model performance comparison dashboard
• Live betting odds comparison for value bet detection

Long-term

• Machine learning model experimentation
• Social features: share models, accuracy leaderboards
• Mobile native apps using Rust shared core
• Public API for third-party integrations