Matrica: AI-Powered Esports Sponsorship Intelligence Platform 🏆

Matrica is an enterprise-grade esports analytics platform that uses a Hybrid AI Architecture to dynamically match global esports teams and brands with top-tier professional players. It fuses complex in-game performance metrics (structured data) with real-world news and brand sentiment (unstructured data).

Currently built for the VALORANT Champions Tour (VCT) 2025.

🚀 Key Features

• Dynamic Data Pipeline (Medallion Architecture): Fully automated ETL pipeline that extracts raw VCT tournament data, cleans it using PySpark, models it into a Star Schema, and uploads it to a production PostgreSQL database.
• Algorithmic Sponsor Matching: Uses PostgreSQL math functions to dynamically calculate Player Popularity, Brand Reputation, and Estimated Budget across hundreds of professionals, matched perfectly against brand slider weights.
• Hybrid AI Analysis:
  • Structured: FastAPI + PostgreSQL (psycopg2) handles real-time ranking algorithms.
  • Unstructured: LangChain + ChromaDB handles RAG (Retrieval-Augmented Generation) on real-world news articles.
  • Generative: Groq LLMs fuse the SQL rankings and ChromaDB context into a coherent, professional summary for sponsors.
• Modern Web Interface: Built with React, Vite, and Tailwind CSS.

🏗️ Architecture

graph TD;
    A[Raw Bronze CSVs] -->|PySpark| B[Silver Cleaned Data];
    B -->|PySpark| C[Gold Star Schema];
    C -->|SQLAlchemy| D[(Neon PostgreSQL)];
    
    E[News Articles JSON] -->|LangChain| F[(ChromaDB Vector Store)];
    
    D -->|SQL Agent| G[FastAPI Backend];
    F -->|RAG Agent| G;
    
    G -->|JSON API| H[React + Tailwind Frontend];

Loading

🛠️ Tech Stack

• Data Engineering: Apache Spark (PySpark), Jupyter, nbconvert
• Database: Neon (Serverless PostgreSQL), ChromaDB (Vector Search)
• Backend: Python, FastAPI, SQLAlchemy, LangChain, Groq LLMs (Llama 3 / Mixtral)
• Frontend: React 18, TypeScript, Tailwind CSS, Vite, Axios

💻 How to Run the Project

1. Run the Data Pipeline (ETL Orchestrator)

The enterprise orchestrator will automatically execute all 6 data pipeline notebooks sequentially, clean the data, and push it to your PostgreSQL database.

# Activate your virtual environment
venv\Scripts\activate

# Run the Orchestrator
python data_pipeline/orchestrator.py

2. Start the Backend API

# In Terminal 1
venv\Scripts\activate
python backend/main.py

API runs on http://localhost:8000

3. Start the Frontend Web App

# In Terminal 2
cd frontend
npm run dev

Website runs on http://localhost:5173

🎥 Project Demo

Check out the full video demonstration of Matrica in action: Watch the Project Demo on Google Drive

---

Built for the VCT 2025 Season. Data modeling and architecture designed for enterprise scalability.