Back to Portfolio

NBA Hall of Fame Predictor

An interactive ML system that estimates Hall of Fame probability from career outcomes, peak performance indicators, and awards profile. The interface lets users query players and inspect model-backed reasoning.

PythonXGBoostNext.jsTypeScriptSports Analytics
LeBron James
Hall of Fame Probability
99.8%
Top Drivers
All-Star selections
Career win shares
MVP and All-NBA awards

Key Performance Indicators

Model Design

Feature Set

Inputs combine volume stats, peak advanced metrics (PER/BPM/VORP), longevity, and accolades to capture both peak and career value.

Gradient Boosted Trees

XGBoost handles nonlinear interactions between awards, usage, and efficiency while remaining interpretable enough for per-player explanations.

Prediction Pipeline

predict_hof.py
def predict_player(player_row):
    features = vectorize_player(player_row)
    probability = float(model.predict_proba(features)[0, 1])
    confidence = confidence_bucket(probability)

    return {
        "player": player_row["name"],
        "hof_probability": probability,
        "confidence": confidence,
        "reasoning": explain_prediction(model, features)
    }