🏏 IPL Live Win Probability Predictor

Ball-by-ball IPL 2nd innings win probability predictor built using Machine Learning (XGBoost) and visualized in Tableau.

---

👨‍💻 Authors

• Om Naik
• Murali Manohara

---

📊 Project Overview

• Dataset: 1,169 IPL matches (2008–2025)
• Records: 278,000+ ball-by-ball deliveries
• Problem Type: Binary Classification (Win / Loss)
• Best Model: XGBoost (Highest Win Recall)

This project predicts live win probability during the 2nd innings of an IPL match using ball-by-ball match conditions.

---

🔧 Tools & Technologies

• Python
  • Pandas
  • NumPy
  • Scikit-learn
  • XGBoost
• Jupyter Notebook
• Tableau Public

---

📈 Model Comparison

Model	Accuracy	Win Recall	False Negatives
Logistic Regression	78%	0.68	4,291
Random Forest	75%	0.69	4,242
XGBoost	77%	0.72 ✅	3,765 (Best)

---

🔍 Key Findings

• Pressure Index is the most important feature (≈52% importance)
• Mumbai Indians and KKR are the strongest chasing teams
• Win probability changes significantly in the last 5 overs
• IPL matches remain unpredictable until the final over

---

📊 Tableau Dashboard

Interactive Dashboard:

👉 https://public.tableau.com/views/IPLLiveWinProbabilityAnalytics/IPLLiveWinProbabilityAnalyticsDashboard?:language=en-US&:sid=&:redirect=auth&:display_count=n&:origin=viz_share_link

Dashboard includes:

• Win Probability Curve
• Matches Won by Team
• Pressure Index Analysis
• Match Summary Statistics

---

📁 Project Files

File	Description
ipl_win_probability_model.ipynb	Machine learning notebook
ipl_win_probability.csv	Ball-by-ball features
ipl_match_summary.csv	Match-level summary
ipl_feature_importance.csv	XGBoost feature importance
ipl_win_probability.twbx	Tableau dashboard

---

🎯 Project Goal

To build a real-time IPL win probability prediction system using machine learning and visualize match dynamics through interactive dashboards.

---

🚀 Future Improvements

• Real-time API integration
• First innings prediction model
• Player-level analysis
• Deep Learning models