I need a web-based system that forecasts three key outcomes—moneyline (H2H), line/spread, and total points (over/under)—for the following sports: NRL, AFL, NBL, NBA, NHL, MLB, and NFL. Core requirements • Model training must rely on historical game data only; I already have access to public datasets and can provide sample CSVs if that helps. • Weather, venue, date, and start-time should be treated as additional features to improve accuracy. • The interface should let me: – browse upcoming fixtures with the model’s probabilities and suggested picks – record the bets I actually place (stake, odds, outcome) – view running profit/loss and basic analytics such as win rate and ROI. Technical notes • Deploy as a responsive web application so I can use it from both desktop and mobile browsers. • I’m comfortable with common stacks (Python + Flask/FastAPI, Node.js, etc.); choose whichever you find most efficient, provided the code is well-commented and easy to maintain. • Please include instructions for retraining the model when new season data becomes available. Deliverables 1. Clean, documented source code for the prediction engine and web front-end. 2. Script or notebook showing data preprocessing, feature engineering, model training, and evaluation metrics. 3. Deployment guide (Docker or similar) so I can spin the app up on my own server. I’m aiming for a solid functional prototype that gets predictions up and running quickly, with room to iterate once real betting results start flowing through the tracker.