Soumya ranjan Padhi

IPL MatchPulse AI – Change One Ball, Rewrite the Match

Cricket fans and analysts can see live scores, but they cannot interactively understand how one ball, one player decision, or one tactical change can alter the entire match outcome. Existing platforms like score apps only show results, not “what-if” scenarios. The problem was to build a system that uses IPL historical data and live match context to predict win probability, explain key turning points, and let users edit any ball, player, or over to instantly see how the result changes. This helps users, students, and analysts better understand match dynamics through data-driven simulation. Process Started by collecting and cleaning IPL ball-by-ball data, then performed exploratory analysis on toss impact, phase-wise scoring (powerplay, middle, death overs), and player performance. Built interactive dashboards to validate insights and identify key winning factors. Designed the core “What-If Match Rewriter” to let users edit any ball, player, or over and instantly re-simulate outcomes. Trained an XGBoost win prediction model using match-state features like score, wickets, run rate, batter impact, and bowler economy. Added AI smart commentary and turning-point detection for explainability. Finally, designed a premium IPL-style responsive UI and tested all scenarios for deployment readiness. Results Built a fully functional AI-powered IPL match simulator capable of live win prediction, what-if scenario simulation, and ball-by-ball match rewriting. Achieved dynamic real-time probability updates, interactive analytics dashboards, and responsive UI across devices. Testing showed accurate match-state recalculations and strong usability through intuitive editing and live commentary features. Next time, I would improve model accuracy further using larger datasets and integrate real live match APIs. Reflection Use more historical data, improve model training with deeper player-level features, and add live API integration for real-time IPL matches.