FIFA Players Salary Prediction & Recommender System

This project leverages the FIFA players dataset to build two powerful tools:

1. Salary Prediction Model: Predicts a player's salary based on their attributes.
2. Player Recommender System: Recommends similar players using feature-based similarity.

---

🚀 Project Overview

This repository contains a machine learning pipeline that:

• Preprocesses the FIFA players dataset.
• Trains regression models to predict player salaries.
• Implements a similarity-based recommendation engine using player features.

The goal is to provide insights into player valuation and help in scouting or strategic team-building decisions.

---

📂 Features

• 📊 Salary Prediction
  Predict a player's salary using features such as age, overall rating, potential, etc.

• 🤝 Player Recommender System
  Suggests similar players based on numerical player attributes using similarity metrics.

---

📌 Getting Started

1. Clone the repository

git clone https://github.com/luisfucros/fifa-players.git
cd fifa-players

2. Run the project locally using Docker Compose

make local-deploy
make local-app

This will:

• Spin up all necessary services: PostgreSQL, MongoDB, Redis, Adminer, MLflow.

• Run the data ingestion pipeline.

• Launch the ML training job.

• Start the FastAPI-based ML application.

3. Access the Services

• MLflow Tracking UI: http://localhost:5000

• FastAPI Docs: http://localhost:8000/docs

• Adminer (DB Admin UI): http://localhost:8080

4. Run Individual Components (Optional)

You can also run specific services manually:

make local-storage            # Start databases (Postgres, MongoDB, Redis, Adminer)
make local-data-ingestion     # Run data ingestion job
make local-mlflow-server      # Start MLflow tracking server
make local-training           # Run the training job
make local-app   

🔮 Future Improvements:

• Add a frontend for easier interaction.

• Include position-based or user/role-aware recommendations.

• Expand to predict player growth potential.