⚽ Fantasy Premier League (FPL) Player Analytics Make a Team

A complete end-to-end data analytics and machine learning project using Fantasy Premier League data to help optimize fantasy team selection, uncover player performance insights, and build predictive models.

---

📥 Data Acquisition & Cleaning

• ✅ Scraping: Data scraped from the official FPL website (7th October 2024) using requests and BeautifulSoup.
• 🧹 Cleaning: Removed duplicates, handled nulls, standardized formats.
• 🏗️ Feature Engineering: Derived new features such as:
  • Points per 90 minutes
  • Attack/Defense efficiency scores
  • Form indicators

---

📊 Analysis & Reporting (Tableau / Power BI)

• 🧠 Created team and player dashboards showcasing:
  • Player rankings by position
  • Team-wise point distribution
  • Position-wise performance heatmaps
• 📌 Built for strategic fantasy decisions and comparisons

---

📈 EDA & Statistical Testing

• 📉 Trend Analysis:
  • Season-wise points, performance over time
  • Position & team-based trends
• 📊 Correlation & Outlier Detection
• 🧪 Statistical Tests:
  • ANOVA: To analyze variance across multiple positions
  • T-Test: To compare top-tier vs mid-tier players
  • Chi-square: For categorical dependencies
  • Mutual Information: For feature relevance

---

🧬 Feature Selection

• 🧪 Filter Methods: Correlation, Variance Threshold
• 🔁 Wrapper Methods: Recursive Feature Elimination (RFE)
• 🧩 Embedded Methods: Feature importance from tree-based models

---

🤖 Modeling

• 🚀 Baseline Models:
  • Linear Regression
  • Decision Tree
  • K-Nearest Neighbors
• 💡 Advanced Modeling (via PyCaret):
  • Random Forest
  • XGBoost
  • Ensemble Models

---

🎯 Hyperparameter Tuning

• Applied GridSearchCV and PyCaret’s AutoML features for optimal parameters.
• Evaluated models using metrics like R², RMSE, MAE across training and validation sets.

---

🌐 Web Development & Hosting

• Built an interactive web app using Streamlit for:
  • Player selection predictions
  • Position-wise comparison
  • Model summary and insights
• [Optional: Add a deployment link if hosted]

---

🔄 Next Steps

• Automate weekly data scraping and model updates
• Add team formation recommender system

---

✅ Conclusion

This project demonstrates a full data pipeline — from scraping raw data to statistical analysis, ML modeling, and visualization. It combines both technical expertise and domain knowledge in football analytics to provide actionable insights.

---

🎯 Who Will Benefit

• 📈 Fantasy Premier League Managers: For smarter weekly picks
• ⚽ Football Enthusiasts: To understand player performance
• 🧪 Data Science Recruiters: Demonstrates end-to-end project lifecycle
• 🏢 Sports Analytics Firms: A strong sample for predictive analytics in sports

---

🔗 Connect with me

• GitHub: github.com/bhumilad
• LinkedIn: linkedin.com/in/lad-bhumi/

---