About Player Data Insights
ML-powered Premier League player prop predictions, built for football fans and researchers who want to go beyond the table.
What is Player Data Insights?
Player Data Insights is a prediction platform that uses machine learning to forecast individual player statistics for Premier League matches. Each gameweek, our models predict whether a player will exceed key performance thresholds - giving you data-driven context before the match kicks off.
How It Works
Data from Sofascore
Player and match statistics are collected from Sofascore, covering every Premier League fixture - lineups, referee assignments, historical performance, and more.
XGBoost & Random Forest models
We train gradient-boosted and random forest models per player prop. Features include recent form, home/away splits, opponent defensive rating, and referee booking tendencies.
Pre-computed nightly
Predictions are generated the night before each gameweek and cached, so the site loads instantly. After results come in, we score each prediction and update our accuracy tracking.
Props We Predict
Accuracy
In GW31 - our first fully live gameweek - the model achieved 89.2% hit rate across 334 position-relevant predictions (3 fixtures; Man City vs Crystal Palace was postponed).
We only score predictions where the player played a meaningful amount of minutes, and we filter props to those relevant for each position - for example, we don't score goalkeeper shot predictions.
Disclaimer
Player Data Insights is for informational and research purposes only. This is not financial advice. Our predictions are generated by ML models trained on historical data and do not take into account bookmaker odds. Past accuracy does not guarantee future results. Always gamble responsibly and within your means.