MotoGP Position Estimator — Manuel Cattoni

MotoGP Position EstimatorA virtual pit wall for MotoGP: live lap-time and position calls, race pace and tyre strategy, read off roughly 73,000 real laps.A machine learning suite acting as a "Virtual Pit Wall": it predicts live lap times, grid positions, race pace and tyre strategy for MotoGP, built on a SQLite pipeline ingesting real PDF timing sheets (2024-2026 seasons, 34 riders, 23 circuits, ~73k laps).

Live Lap Time & Position Predictor: estimates a rider's final lap time and live leaderboard position at each sector, using an ensemble of 5 regressors (XGBoost, SVR, Neural Network, Ridge, AdaBoost).

Race Pace & Tyre Strategy Engine: predicts average race pace, tyre degradation and likely compound choice from Free Practice telemetry and weather conditions.

Race Weekend Analysis (new): per-session report card, an "ideal lap" execution score, and lap-by-lap position/pace charts.

Analytics Dashboard: algorithmic rider ratings, momentum indicator, circuit breakdown, head-to-head comparison, team performance trends.

Advanced EDA: 3D environmental mapping, Random Forest feature importance, K-Means/PCA circuit clustering.

Tech stack: Python, Flask, scikit-learn, XGBoost, Pandas, SQLite | RenderA project by Manuel Cattoni.