🥋 UFC Fight Winner Prediction (Machine Learning)
A data science project that predicts the winner of a UFC fight using real fighter statistics and historical fight data. This project includes web scraping, feature engineering, ML modeling, and a Streamlit-based UI.
📌 Project Features ✔ Web Scraping
Scraped data from ufcstats.com, including:
Fighter attributes (height, reach, stance, age, SLpM, StrAcc, SApM, StrDef, TDAvg, TDAcc, TDDef, SubAvg)
Full UFC fight history with results and methods
✔ Data Cleaning & Feature Engineering
Converted height/reach to cm
Extracted wins/losses/draws/NC from fighter records
Calculated age, win rate, finish rate
Created delta stats (Fighter1 − Fighter2) used for prediction:
striking deltas
grappling deltas
physical deltas
performance deltas
✔ Machine Learning
Model: RandomForestClassifier
Scaler: StandardScaler
Trained on engineered matchup dataset
Predicts winner + confidence score
✔ Streamlit UI
UFC-themed layout (Red Corner vs Blue Corner)
Fighter selection dropdowns
Fighter stat cards
Delta feature table
Winner prediction with confidence
🛠 Tech Stack
Python · Pandas · NumPy · BeautifulSoup · Scikit-learn · RandomForest · Streamlit
frontend -
