The world's first browser-based ML combat system where algorithms battle using real performance metrics.

Choose your ML fighters and watch them train on real data!

Upload your CSV dataset and watch machine learning algorithms battle Pokemon-style! Each algorithm trains for real using scikit-learn in your browser, then fights based on authentic performance metrics like accuracy, precision, and recall.

🌳 Random Forest vs 🧠 Neural Network - who wins on YOUR data?

Drag & drop your CSV with Pokemon-style "Wild Dataset Appeared!" interface

• 🔥 Real ML Training - Actual scikit-learn algorithms running in browser via Pyodide
• 📊 Live Analytics - Real-time training progress with interactive charts and dashboards
• 🎮 Pokemon Battle System - Nostalgic interface with sprites, animations, and type advantages
• 📈 Performance Visualization - Radar charts, rankings, and detailed metric comparisons
• 🧠 Educational AI - Gemini API provides dataset insights and learning explanations
• 📱 Responsive Design - Works perfectly on desktop and mobile

• Node.js 18+
• Modern browser with WebAssembly support

# Clone the repository
git clone https://github.com/Rayane0001/predictive-combat-arena.git
cd predictive-combat-arena

# Install dependencies
npm install

# Start development server
npm run dev

Create a .env file:

VITE_GEMINI_API_KEY=your_gemini_api_key_here

1. 📁 Upload Dataset - Drag & drop your CSV file
2. 🔍 AI Analysis - Get instant insights about your data
3. ⚔️ Choose Fighters - Select two ML algorithms
4. 🏃‍♂️ Watch Training - See real-time progress as algorithms train
5. 📊 View Analytics - Explore performance charts and metrics
6. ⚡ Epic Battle - Pokemon-style combat based on real ML performance!

Configure your battle mode and AI difficulty before the epic showdown

Experience authentic Pokemon combat with real ML performance determining the outcome!

Watch algorithms battle with sprites, animations, and authentic Pokemon mechanics

• SvelteKit ^2.22.0 - Modern web framework with SSR
• Svelte ^5.0.0 - Reactive component framework
• TypeScript ^5.0.0 - Type-safe JavaScript
• Vite ^7.0.4 - Lightning-fast build tool

• Pyodide ^0.28.0 - Python runtime in WebAssembly
• scikit-learn (via Pyodide) - ML algorithms and training

• D3.js ^7.9.0 - Dynamic data visualizations
• Papa Parse ^5.5.3 - CSV parsing and processing

• Tailwind CSS ^4.0.0 - Utility-first CSS framework
• @tailwindcss/typography ^0.5.16 - Typography plugin
• Lucide Svelte ^0.525.0 - Icon library
• Custom CSS Animations - Pokemon-style battle effects

• Google Gemini API - Dataset analysis and insights

• ESLint ^9.18.0 - Code linting
• Prettier ^3.4.2 - Code formatting
• Vitest ^3.2.3 - Unit testing
• Playwright ^1.53.0 - E2E testing

Combat stats are calculated from real ML performance:

• Attack = Precision × 100
• Defense = Recall × 100
• Speed = 100 - (Training Time × 10)
• Health = Accuracy × 100 + Base Robustness
• Critical Hit Rate = F1-Score

Learn ML concepts through gameplay:

• Algorithm Comparison - See real performance differences
• Metric Understanding - Visualize precision vs recall trade-offs
• Dataset Analysis - Understand what makes data challenging
• Interactive Learning - Hover tooltips explain ML concepts
• Performance Visualization - Charts show training progress

Built for Data Hackfest 2025 - 48 hours of intensive development combining education, entertainment, and technical innovation.

• Rayane Rousseau - Lead Developer
• Gourav Sharma - Member
• Rohan Fernandez - Member

We welcome contributions!

# Install dependencies
npm install

# Run development server
npm run dev

# Build for production
npm run build

# Run tests
npm run test

• Pokemon - For the inspiration and nostalgic aesthetics
• scikit-learn - For making ML accessible and powerful
• Pyodide - For bringing Python ML to the browser
• SvelteKit - For the amazing developer experience
• Google Gemini - For intelligent dataset analysis

• Video Demo: YouTube Demo

Making machine learning accessible, one Pokemon battle at a time! 🚀

Built with ❤️ during Data Hackfest 2025