🏏 Batting Through Numbers: A Data-Driven Analysis of Cricket

Welcome to our repository where data meets passion in the realm of cricket! This project is a detailed exploration of cricket analytics, where we use data science to uncover hidden patterns and optimize strategies.

🌟 Project Overview

"Batting Through Numbers" is not just about cricket analytics; it's a blend of our love for the game and statistical analysis. We delve deep into datasets to discover what lies beneath the surface of scorecards and player statistics.

🎯 Research Questions Addressed

1. Association Between Batting Position and Dismissal:

   • We investigate how a player’s batting position correlates with their likelihood of getting dismissed, providing insights that could influence batting order decisions.

2. Distribution of Dismissal Types for Top-Scoring Batsmen:

   • This analysis focuses on understanding how top performers typically get out, which may help coaches tailor specific training to mitigate these dismissals.

3. Forecasting Batting Performance Using Machine Learning:

   • By employing historical data on similar batting positions, we develop predictive models to forecast batting performances, offering a tool for strategy formulation.

🎯 Objectives

• Analyzing T20 Ball-by-Ball Data: Using detailed ESPN datasets, we uncover complex play patterns and performance metrics.
• Predictive Modeling: We build models to predict match outcomes for the LUMS cricket team and other international teams, leveraging historical performance data.
• Cricket Storytelling: Beyond numbers, we narrate stories of cricket matches, offering insights that are often overlooked.

📂 Repository Structure

• models: Contains machine learning models saved as .pkl files for various international cricket teams.
• data: Includes datasets such as:
  • ball_by_ball_it20(train dataset).csv: Detailed ball-by-ball play data for training our models.
  • LUMS Cricket Dataset - Batting.csv: Specific dataset for batting performance of the LUMS team.
  • LUMS Cricket Dataset - Balling.csv: Specific dataset for bowling performance of the LUMS team.
• CricketDataAnalysis.ipynb: Jupyter notebook containing all the analysis, from data preprocessing to model training and evaluation.

🔧 Tools & Technologies Used

• Python: For all computing and data analysis.
• Pandas & NumPy: For data manipulation.
• Matplotlib & Seaborn: For data visualization.
• Scikit-Learn: For predictive modeling and machine learning.
• Jupyter Notebook: For interactive code development and testing.

🔧 Data Preparation and Transformation

Data Creation

• Data Prep:

  • Loaded, cleaned, and simplified batting/bowling datasets.
  • Handled null values and dropped unnecessary columns.

• Data Transformation:

  • Developed specific functions for dataset generation.
  • Imputed numeric nulls, mapped categorical data, and standardized columns.

• Summarized Data:

  • Condensed ball-by-ball details into summarized player stats per match.
  • Key metrics include runs, balls faced, batting position, etc.

• Result:

  • Created a concise dataset focusing on player performance in matches.
  • Columns include: batter, runs, ballsPlayed, howOut, battingPosition, teamName, batFirst, targetScore, extraRuns, winner, matchId.

📈 Statistical Analysis, Results, and Key Findings

Statistical Test

• Chi-square statistic: 324.20
• P-value: 3.84e-62

Average Treatment Effect (ATE)

• ATE: 0.177

• Inference: Top-scoring batsmen exhibit a higher average dismissal type compared to non-top-scorers.

• Statistical Analysis: We performed extensive statistical tests, like the Chi-squared test, to explore the association between a player's batting position and their likelihood of dismissal.

• Predictive Performance: Our models, trained on diverse datasets, predict outcomes with high accuracy, demonstrated by metrics like RMSE (Root Mean Squared Error).

• Strategic Insights: Recommendations on batting order and player positioning that can significantly affect match outcomes.

🚀 How to Use This Repository

1. Clone the repository:

   git clone https://github.com/yourusername/CricketDataAnalysis.git
   

2. Install the necessary Python packages:

    pip install -r requirements.txt
   

3. Run the Jupyter notebooks:

    jupyter notebook
   

📝 Conclusion

Our project bridges the gap between raw data and actionable cricket insights. It demonstrates how data-driven strategies can be applied to enhance understanding and performance in sports analytics.

📚 Additional Resources

• Final Project Blog: Complete insights into the project. Read here

• EDA Blog: Deep dive into the exploratory data analysis. Read here

🤝 Contributing

Feel free to fork this repo, suggest changes via pull requests, or discuss ideas in issues. Your contributions are always welcome!

🌐 Contact

For any further questions, please email us at [email protected], or raise an issue in this repository.

⭐ Support

If you find this repository helpful, please consider giving it a star to help others find it!