Profitable EPL Betting Algorithm

I want to turn raw English Premier League data into a bookmaker-grade model that highlights value on the 1X2, BTTS, and Over/Under markets before kick-off. Pre-match odds evaluation will be the core signal, but I’m happy to blend in any additional statistics you can justify.

I already hold some historical EPL results and closing prices, yet I’m equally open to tapping into paid APIs or publicly scraped feeds if you can point me to better, cleaner, or more granular sources.

Here is what I need from you:
• An automated data pipeline that fetches, cleans, and stores the chosen datasets.
• A modelling framework in Python or R (Pandas, scikit-learn, XGBoost, or comparable libraries) that outputs implied probabilities and fair odds for each of the three target markets.
• Robust back-testing with walk-forward validation, showing historical yield, hit-rate, and maximum drawdown.
• A simple staking or Kelly-based optimiser so we can simulate real bankroll performance.
• Clear documentation plus reproducible code so I can rerun the process each matchday and later extend it to in-play.

If the model can demonstrate consistent positive ROI after realistic bookmaker margins and transaction costs, I’ll green-light the next phase.