AI Predictive Model for MLB Games
Budget: $30 – $250 USD
*** This will run on a Raspberry Pi 5 RasbianOS ***
I want to build an AI model that accurately predicts the winner of MLB games. The system should automatically gather data for the 2025 MLB season using available APIs and official data sources such as Baseball Reference, MLB.com, and others, following all legal and licensing guidelines. The method of data retrieval doesn’t matter as long as it’s reliable, comprehensive, and fully compliant.
The system will use this data—covering past, current, and upcoming games and all relevant team and player stats—to train and continuously retrain the AI model, ensuring predictions improve as the season progresses. For each game, the model should generate a prediction showing the likely winner, the confidence level (e.g., 86%, 40%), and a brief explanation of the reasoning so users can understand why the model made its pick.
I want a user-friendly web interface with three main sections: past games, today’s games, and future games. Each section should display results in a clean, sortable table format, listing game date, teams, predicted winner, confidence, and explanation. There should be easy navigation between sections and a clear way to return to the home page.
For today’s games, the system should automatically display all matchups scheduled for the current day, with start times in Central Standard Time. Future games should show predictions based on the latest available data. The entire system should update itself automatically on a schedule (using tools like cron or background services), always working with the most current stats and results.
Overall, I want a hands-off, fully automated solution that uses legal API data sources, retrains the AI, and presents up-to-date MLB predictions in a web app, with no manual intervention required.
I want to build an AI model that accurately predicts the winner of MLB games. The system should automatically gather data for the 2025 MLB season using available APIs and official data sources such as Baseball Reference, MLB.com, and others, following all legal and licensing guidelines. The method of data retrieval doesn’t matter as long as it’s reliable, comprehensive, and fully compliant.
The system will use this data—covering past, current, and upcoming games and all relevant team and player stats—to train and continuously retrain the AI model, ensuring predictions improve as the season progresses. For each game, the model should generate a prediction showing the likely winner, the confidence level (e.g., 86%, 40%), and a brief explanation of the reasoning so users can understand why the model made its pick.
I want a user-friendly web interface with three main sections: past games, today’s games, and future games. Each section should display results in a clean, sortable table format, listing game date, teams, predicted winner, confidence, and explanation. There should be easy navigation between sections and a clear way to return to the home page.
For today’s games, the system should automatically display all matchups scheduled for the current day, with start times in Central Standard Time. Future games should show predictions based on the latest available data. The entire system should update itself automatically on a schedule (using tools like cron or background services), always working with the most current stats and results.
Overall, I want a hands-off, fully automated solution that uses legal API data sources, retrains the AI, and presents up-to-date MLB predictions in a web app, with no manual intervention required.