Soccer Player Market Value Scout

A full-stack machine-learning service that predicts the market value of soccer players from the top-5 European leagues and explains what drives each prediction.

• Live web app (Streamlit): https://418final-uxdflcbiixufvv9rpxytej.streamlit.app/
• Prediction API (Cloud Run): https://market-value-api-348858993647.us-central1.run.app/docs
• Code: github.com/lacygus/418_final

What it does. Pick any of 2,013 players and see the model’s predicted value, an 80% prediction interval, and per-feature contributions that explain it. Browse the most over- and under-valued players, compare leagues and positions, and inspect model accuracy league-by-league.

Stack. RandomForestRegressor on log(market_value) (test R² = 0.59), a FastAPI service containerized and deployed to Google Cloud Run, and a Streamlit front end, with the data scraped and cleaned from Transfermarkt and Sofifa.

Python · scikit-learn · FastAPI · Streamlit · Docker · Google Cloud Run