🏏 T20 Cricket Data Analysis Project

📊 End-to-End Data Analytics Pipeline (Python · Web Scraping · SQL · Power BI)

---

🧩 Overview

This project is a complete end-to-end Cricket Analytics workflow where T20 player's batting & bowling performance data is collected through web scraping, cleaned using Python and transformed the json -> csv file using Pandas, and visualized through an interactive Power BI dashboard.

The goal is to identify player performance trends, strengths/weaknesses, and build a data-driven Dream T20 Playing XI using role-based parameter criteria.

Inspired by: 🎥 Codebasics Cricket Analytics Series
Repository: Cricket Data Analysis

---

🗂️ Dataset

The dataset consists of T20 batting and bowling statistics scraped from ESPN Cricinfo.

Data Includes:

• Player innings-level batting data
• Bowling performance summary
• Player role metadata
• Team-wise and match-wise T20 stats

Data Size: ~2,500+ records
Format: CSV + cleaned DataFrames

---

🔧 Tools & Technologies

Tool	Purpose
Python (Pandas, BeautifulSoup, Requests)	Web scraping, cleaning, preprocessing
Jupyter Notebook	EDA & feature engineering
PostgreSQL	SQL-based filtering & analysis
Power BI	Dashboard creation & data storytelling
DAX Measures & Calculated Columns	KPI calculations
Parameter Scoping PDFs	Role-based player criteria

---

🕸️ Step 1: Web Scraping (Python)

Using BeautifulSoup and Requests, raw player data was extracted from ESPN Cricinfo.

Scraped Features:

• Batting average, strike rate, boundary %
• Bowling economy, strike rate, dot ball %
• Player roles & innings breakdown

Challenges solved:

• HTML inconsistencies
• Pagination scraping
• Merging batting + bowling data

(Notebook reference: t20_data_preprocessing.ipynb)

---

🧹 Step 2: Data Cleaning & Preprocessing

Performed using Pandas in Jupyter Notebook.

Tasks Completed:

• Removed duplicates & missing values
• Standardized columns
• Calculated new metrics:
  • Boundary %
  • Balls faced per innings
  • Dot ball %
  • Bowling strike rate
• Merged player role information

Generated output files:

• batting_summary_cleaned.csv
• bowling_summary_cleaned.csv

---

🧮 Step 3: Applying Cricket Selection Parameters

Role-based filtering performed using criteria from your Parameter Scoping PDFs.

🔹 Openers Criteria

• Avg > 30
• SR > 140
• Boundary % > 50%
• Position < 4

🔹 Anchors / Middle Order

• Avg > 40
• SR > 125
• Balls faced > 20
• Position > 2

🔹 Finishers

• SR > 130
• Balls faced > 12
• Position > 4

🔹 All-Rounders

• Bat SR > 140
• Economy < 7
• Bowling strike rate < 20

🔹 Specialist Fast Bowlers

• Dot % > 40
• Economy < 7
• Bowling strike rate < 16
  These filters were applied using Pandas (for data filtering, feature engineering, and transformation) and Power BI DAX (for KPIs inside the dashboard).

---

🧮 Step 4: Player Filtering & Role Classification (Pandas + DAX)

After preprocessing the batting and bowling datasets, all player selection logic was implemented using Pandas inside the Jupyter Notebook.

🔍 Role-Based Filtering Performed

• Filtered Openers using Avg > 30, SR > 140, Boundary% > 50 and Position < 4
• Identified Anchors using Avg > 40, SR > 125, Balls Faced > 20
• Selected Finishers based on SR > 130 and lower-order batting position
• Extracted All-rounders using combined batting & bowling metrics
• Ranked Fast Bowlers using Economy < 7, Dot% > 40, Strike Rate < 16

🧠 Feature Engineering with Pandas

• Calculated Boundary%, Dot Ball%, Balls Faced per Innings, Bowling Strike Rate, Runs per Over, etc.
• Merged batting & bowling datasets for all-rounder evaluation
• Applied filtering conditions to create final role-wise player lists

📈 Additional DAX Logic Used in Dashboard

Inside Power BI:

• Created KPI cards using DAX (Avg Runs, Strike Rate, Economy, Dot %, Wickets)
• Built ranking measures for comparing players within roles
• Designed slicer-based dynamic comparisons

---

📊 Step 5: Power BI Dashboard

The dashboard includes:

• 🏏 Player role filters
• 📈 KPI cards (Runs, Avg, SR, Economy, Dot %, Wickets)
• 🔥 Best Playing XI (Data-driven)
• 📌 Player role comparison
• ⚡ Fully interactive slicers (team, position, role)
• 🎨 Professional theme based on dashboarding principles

Dashboard File:
📁 Cricket_data_analysis.pbix

Reference Files:

• DAX-Measures-and-Calculated-Columns.xlsx
• Dashboarding-Tips.pdf
• Parameter-Scoping.pdf

---

🧠 Key Insights

• Identified openers with 150+ Strike Rate and high boundary %.
• Found fast bowlers with Economy < 7 and Dot ball % > 40%.
• Anchors with Average > 40 and high balls-per-innings stability.
• Built a Data-driven Best T20 Playing XI.
• Highlighted role-specific strengths & weaknesses.

---

⚙️ How to Run the Project

1️⃣ Install Dependencies

Make sure you have the required Python libraries:

2️⃣ Run the Web Scraping Script

This script collects raw player data from ESPN Cricinfo.

3️⃣ Preprocess Data (Cleaning + Feature Engineering)

Open the Jupyter Notebook:

 t20_data_preprocessing.ipynb

Inside this notebook, the following steps are performed:

• Load scraped JSON data
• Clean & preprocess datasets using Pandas
• Create derived metrics such as:
  • Boundary%
  • Dot Ball%
  • Average Balls Faced
  • Bowling Strike Rate
  • Bowling Economy
• Merge batting & bowling data where required
• Export final cleaned CSV files for dashboard use

---

4️⃣ Export Clean CSV Files

After preprocessing, the notebook generates the following cleaned files:

batting_summary_cleaned.csv
bowling_summary_cleaned.csv

These are the input data sources for the dashboard.

5️⃣ Open the Power BI Dashboard

Open the dashboard file:

Cricket_data_analysis.pbix

Inside Power BI, you can:

• Explore player filters (teams, roles, positions)
• View KPI cards (SR, Avg, Economy, Dot%, Wickets)
• Analyze role-based player rankings
• Examine department-wise performance
• Check the final Data-Driven T20 Best Playing XI

📚 Reference

Project inspired by:

🎥 Codebasics Cricket T20 Analytics

👨‍💻 Author

Utkarsh Singh

GitHub: https://github.com/Utkarshv05

Email: utkarshsingh2504@gmail.com

LinkedIn: https://www.linkedin.com/in/utkarsh-singh-us02052004/