Soccer Analytics with Machine Learning. Learning Predictive Modeling Techniques with Sports Data Haipeng Gao, Ari Joury, Weining Shen

Dostępność produktu

Produkt został zweryfikowany pod kątem ułatwień dostępu więcej przeczytasz w deklaracji dostępności.

Spis treści książki

• Preface
  • Who Should Read This Book
  • Why We Wrote This Book
  • Navigating This Book
  • Conventions Used in This Book
  • Using Code Examples
  • OReilly Online Learning
  • How to Contact Us
  • Acknowledgments
• 1. The Soccer Analytics Landscape
  • The Problems That Soccer Analytics Helps Solve
    • Soccer Clubs: Making Decisions Across the Organization
      • Match preparation and tactical decisions
      • Player recruitment and roster management
      • Player performance and development
      • Club operations and fan engagement
    • Betting Markets: Modeling Uncertainty and Pricing Outcomes
    • Fans and Analysts: Reading the Game Beyond the Scoreline
  • The Data-Driven Evolution in Soccer
    • From Outcomes to Process: The Limits of Traditional Analysis
    • The Data Revolution: From Event Logs to Tracking Systems
    • Analytics in Practice: Metrics, Applications, and Scope
    • Why Soccer Resists Easy Analysis
  • A Framework for Soccer Analysis
    • The Three Types of Analysis
    • Data Sources for Modern Analysis
    • From Player to League: Levels of Aggregation
    • Principles That Run Through Everything
  • Looking Ahead: A Guided Tour of the Book
  • Conclusion
• 2. Python Fundamentals: Building Your Analytics Toolkit
  • Python: The Language Behind Modern Analytics
  • Your Analytics Workspace: Setting Up the Environment
    • Installation: Getting Python Ready
    • Virtual Environments: Isolating Your Projects
    • Jupyter Notebook: Creating a Workspace for Interactive Analysis and More
      • Features and benefits of Jupyter notebooks
      • Your first notebook: Getting up and running
    • Repositories: Storing Your Project Files
  • Python Basics: The Core Moves
    • Primitive Data Types: Representing Basic Information
    • Basic Operations: Working with Values and Comparisons
    • Variable Assignment: Storing and Reusing Information
    • Data Structures: Organizing Information
      • Lists: Ordered and flexible collections
      • Tuples: Fixed records and immutable structure
      • Sets: Collections of unique elements
      • Dictionaries: Key-value organization
    • Conditional Statements: Writing Decision Rules
    • Loops: Repeating Computation Efficiently
    • List Comprehensions: Performing Compact and Pythonic Iteration
    • Functions: Providing Reusable Logic for Analysis
  • Looking Ahead: From Python Basics to Real Analysis
  • Conclusion
• 3. Exploratory Data Analysis in Soccer
  • Understanding Soccer Data Levels
    • Match-Level Data
    • Player-Level Data
    • Event-Level Data
  • What Is EDA in Soccer Analytics?
  • Loading StatsBomb Soccer Data
    • Getting the Data
    • Loading the Data
    • Taking a First Look at the Data
  • Structured EDA by Data Types
    • Numerical Features: The Quantifiable Game
      • Continuous variables
      • Discrete variables
    • Categorical Features: The Qualitative Aspects
    • Ordinal Features: The Ordered Categories
  • Preliminary Exploration: Getting Your Bearings
    • Building Tournament DataFrames
    • Counting the Frequency of Event Types
    • Filtering: Passes by Player or Team
    • Grouping: Per-Match Pass Summaries
    • Aggregating: Shots, Goals, and xG by Team
  • Basic Data Cleaning for Soccer
    • Handling Missing Values
    • Standardizing Names
  • Visualizing Soccer Data
    • A Quick Note on Matplotlib and Seaborn
    • Bar Charts and Rankings
    • Histograms and Distributions
    • Boxplots and Variability
    • Time Series and Cumulative Plots
    • Scatterplots: Shot Volume Versus Goals
  • Soccer-Specific Visualizations with mplsoccer
    • Pass Location Heatmaps
    • Advanced Pitch Visualizations: Passing Networks
    • Basic Pitch Visualizations
    • Advanced Heatmap Visualizations
  • Case Study: Japan Womens at the 2019 World Cup
    • Team-Level Overview
    • Match Flow
    • Spatial Patterns
    • Player-Level Insights
    • Wrap-up of the Case Study
  • From EDA to Machine Learning
    • Features as Building Blocks
    • The Standard Pipeline
  • Looking Ahead: From Exploration to Regression
  • Conclusion
• 4. Regression Techniques for Soccer Analytics
  • From Lines on a Field to Lines of Best Fit: Linear Regression
    • Predicting Market Value: A Players Worth
    • Building the Model: The Line of Best Fit
    • Visualizing the Model
    • Evaluating Performance: How Good Is Our Model?
  • Counting Goals: Poisson Regression
    • Predicting Goals in a Match
    • Visualizing the Poisson Model
  • Finding Similar Players: K-Nearest Neighbors Regression
    • Predicting Player Value with KNN
    • Choosing the Right k
  • Evaluating and Interpreting Regression Models in Soccer Contexts
    • Using Key Performance Metrics
    • Understanding the Importance of a Validation Set
    • Interpreting Coefficients
    • Visualizing KNN: Decision Boundaries
    • Comparing Models
  • Advanced Topics in Poisson Regression: Overdispersion and Zero-Inflation
    • Overdispersion: When Variance Exceeds the Mean
    • Negative Binomial Regression
    • Zero Inflation: Too Many Zeros
    • Hurdle Models: An Alternative Approach
  • Model Diagnostics: Ensuring That Your Model Is Sound
    • Residual Analysis: Looking for Patterns
    • Goodness-of-Fit Tests
    • Influential Observations: Outliers that Matter
    • Model Selection: Choosing the Best Model
  • Practical Case Study: Predicting Match Outcomes
    • Generating the Dataset
    • Performing Exploratory Analysis
    • Building and Comparing Models
    • Evaluating on Test Data
    • Interpreting the Results
  • Feature Engineering for Better Predictions
    • Interaction Features
    • Polynomial Features
    • Categorical Features
    • Lagged Features
    • Domain-Specific Features
  • Common Pitfalls and Practical Tips
    • Pitfall 1: Overfitting to the Training Data
    • Pitfall 2: Ignoring Multicollinearity
    • Pitfall 3: Using the Wrong Model for the Data
    • Pitfall 4: Forgetting About Data Quality
    • Tip 1: Start Simple, Then Add Complexity
    • Tip 2: Visualize, Visualize, Visualize
    • Tip 3: Document Your Process
    • Tip 4: Validate with Domain Experts
  • Looking Ahead: Where Regression Meets the Real World
    • From Regression to Classification
    • Ensemble Methods: Combining Models
    • Deep Learning: The Frontier
    • Time Series and Sequential Data
    • Causal Inference: Beyond Prediction
    • The Human Element
  • Conclusion
• 5. Predicting Soccer Outcomes with Classification
  • From Coin Flips to Goal Predictions: Understanding Classification
  • Logistic Regression: Your First Classifier
    • The Intuition: From Numbers to Probabilities
    • Hands-On: Building an Expected Goals Model
  • Measuring Success: How Good Are Your Predictions?
    • The Confusion Matrix: A Report Card for Your Model
    • Precision and Recall: The Two Sides of Performance
    • The ROC Curve: A Complete Performance Picture
    • Evaluation of Probabilistic Models with Cost Functions
      • Log loss (cross-entropy loss)
      • Brier score
      • Calibration
  • Beyond Binary: Multiclass Classification
  • K-Nearest Neighbors: Finding Similar Situations
    • The Intuition: Birds of a Feather
    • Hands-On: KNN for Shot Prediction
    • Choosing the Right K
  • Case Study: Building a Complete Match Outcome Predictor
    • Step 1: Define the Problem
    • Step 2: Gather and Prepare the Data
    • Step 3: Train and Evaluate the Model
    • Step 4: Interpret the Results
    • Key Takeaways from the Case Study
  • Practical Tips and Common Pitfalls
    • Handling Imbalanced Data
    • Selecting and Engineering Features
    • Avoiding Overfitting
    • Choosing When to Use Which Algorithm
  • Tuning the Decision Threshold: Finding the Sweet Spot
    • Understanding the Threshold Trade-Off
    • Finding the Optimal Threshold
    • Deciding When to Adjust the Threshold
  • Comparing Models: Which Classifier Is Best?
    • The Model Comparison Framework
    • Hands-On: Model Comparison for xG
    • Visualization of Model Performance
  • Advanced Topics: Going Deeper
    • Scaling and Normalizing Features
    • Handling Missing Data
    • Combining Classifiers with Ensemble Methods
  • Real-World Considerations
    • Data Quality and Availability
    • Model Maintenance and Updates
    • Ethical Considerations
  • Looking Ahead: From Linear Boundaries to Decision Trees
  • Conclusion
• 6. Advanced Classification Methods
  • The Classification Tree: Soccers Decision Flowchart
    • The Intuition: Learning the If-Then Rules of Soccer
    • Hands-On: Building a Tree to Predict Pass Outcome
    • Interpreting the Soccer Intelligence
    • Understanding Pros and Cons: When to Use a Single Tree
    • Tuning the Tree: Preventing Overfitting
  • Random Forest: The Wisdom of the Crowd
    • The Intuition: Ensemble Learning
    • Hands-On: Building a Random Forest for to Predict Match Outcome
    • Interpreting Feature Importance
    • Tuning Random Forests
    • Using Out-of-Bag Error
    • Understanding When Random Forests Struggle
  • XGBoost: Learning from Mistakes
    • The Intuition: Sequential Learning
    • Hands-On: XGBoost for xG Prediction
    • Performing SHAP Analysis: Explaining the Opaque Model
    • Tuning XGBoost: The Art of Hyperparameter Optimization
    • Knowing When to Use XGBoost Versus Random Forest
  • Practical Applications: From Theory to Practice
    • Example 1: Building a Comprehensive xG Model
    • Example 2: Predicting Player Actions
    • Example 3: Recognizing Tactical Patterns
    • Handling Imbalanced Data in Soccer
    • Using Feature Engineering for Tree-Based Models
  • Looking Ahead: From Handcrafted Rules to Learned Representations
  • Conclusion
• 7. Deep Learning for Soccer Analytics: From Intuition to Implementation
  • Why Deep Learning for Soccer? Beyond Traditional Models
    • Looking Inside the Model: How Neural Networks Capture Patterns
      • Neurons
      • Activation functions
        • The ReLU
        • The sigmoid function
        • The hyperbolic tangent (tanh)
      • Forward propagation
      • Loss functions
    • Learning from Data: Training, Loss, and Optimization
      • Backpropagation
      • Gradient descent
      • Regularization techniques
  • Building Your First Soccer Prediction Model with PyTorch
    • Binary Outcomes: Will the Home Team Win?
      • Preparing the data
      • Defining the model architecture
      • Training the model
      • Evaluating the model
    • Match Results: Win, Draw, or Loss
      • Preparing the data
      • Defining the model architecture
      • Training the model
      • Evaluating the model
    • Goal Prediction: Modeling Scoring Output
      • Defining the model architecture
      • Training the model
      • Interpreting the predictions
    • Full Match Simulation: From Goals to Outcomes
      • Using multitask learning
      • Preparing the data
      • Defining the model architecture
      • Training the model
      • Understanding why this matters
  • A Second Lens: Implementing the Model in TensorFlow
    • Binary Outcomes: Will the Home Team Win?
      • Model architecture
      • Training and evaluation
    • Match Results: Win, Draw, or Loss
      • Model architecture
      • Training and evaluation
    • Goal Prediction: Modeling Scoring Output
      • Model architecture
      • Training and evaluation
    • Full Match Simulation: From Goals to Outcomes
      • Model architecture
      • Training and evaluation
  • Looking Ahead: From Flexible Models to Smarter Inputs
  • Conclusion
• 8. Feature Engineering for Soccer Analytics
  • Feature Engineering: From Raw Data to Predictive Signals
  • Performance Features: The Fundamentals
    • Goals For and Goals Against: Measuring Scoring Performance
    • Venue Splits: Considering Home Versus Away
    • Shot Quality: Evaluating Chance Creation
  • Tactical Features: Describing How Teams Play
    • Passing and Possession: Measuring Ball Control
    • Tackles, Interceptions, and Recoveries: Tracking Defensive Actions
    • Pressing Intensity: Applying Defensive Pressure
    • Set Pieces: Capturing Dead-Ball Situations
  • Contextual Features: Incorporating Match Conditions
    • Rest and Scheduling: Modeling Fatigue and Congestion
    • Form and Momentum: Capturing Recent Performance Trends
  • Ranking Features and Advanced Ranking Algorithms: Estimating Team Strength
    • The Colley Matrix: A Linear Algebra Approach
    • PageRank: A Network-Based Perspective
    • Elo Ratings: A Dynamic Rating System
    • A Comparison of Methods: Strengths and Trade-Offs
  • Player-Level Features
  • Feature Engineering Pipeline: From Data to Model Input
  • Case Study: Predicting the 2019 Womens World Cup Final
  • Looking Ahead: From Better Inputs to Harder Decisions
  • Conclusion
• 9. The Art and Science of Soccer Betting
  • Soccer Betting Basics
    • Types of Soccer Bets
      • Match result
      • Over/under
      • Both teams to score
      • Correct score
    • Odds
      • Decimal odds (European)
      • Fractional odds (UK)
      • Moneyline odds (American)
      • Implied probabilities
      • The overround
    • Arbitrage Opportunities
      • Understanding the basic arbitrage formula
      • Identifying arbitrage opportunities
      • Calculating optimal stakes if arbitrage exists
  • The Real Dataset Used in This Chapter
  • Choosing a Side
    • Always Pick the Home Team
    • Always Pick the Lowest Odds
    • Use a Simple Machine Learning Baseline
  • Betting Strategies
    • Fixed-Stake Betting
      • Fixed stakes on the home-team rule
      • Fixed stakes on the lowest-odds rule
    • Kelly Criterion
      • Kelly on the home-team rule
      • Kelly on the lowest-odds rule
  • From Predictions to Potential Profit
    • The Reality of Profitability
    • Practical Considerations
      • Bankroll management
      • Record keeping and the closing line value
      • Psychology and losing streaks
      • Shopping for odds
    • The Importance of Discipline
    • The Role of Technology
    • Ethical Considerations
  • Looking Ahead: Building on Top of the Fundamentals
  • Conclusion
• 10. Whats Next for Soccer Analytics?
  • What This Book Was Really About
    • From Questions to Data
    • Data Preparation Is Part of the Analysis
    • Exploration Before Prediction
    • Feature Engineering Is Where Domain Knowledge Enters
    • Models Support Judgment, They Do Not Replace It
  • The Next Stage in Soccer Analytics
    • From Static Datasets to Living Data Pipelines
    • From Team-Level Models to Player-Level and Lineup-Level Models
    • Richer Data: Tracking, Video, and Multimodal Analysis
    • Stronger Modeling Directions
    • From Notebooks to Analytical Products
    • AI and Large Language Models in the Analytics Workflow
  • Challenges and Limits
    • Data Quality and Access
    • Interpretability and Trust
    • Uncertainty Will Not Disappear
    • Ethics, Privacy, and Responsible Use
  • Beyond Soccer
  • A Practical Roadmap After This Book
  • Final Thoughts
• A. References and Further Reading
  • Further Reading by Topic
    • Foundational Statistics and Regression
    • Advanced Statistical Methods
    • Machine Learning and Deep Learning
    • Sports Analytics
    • Soccer-Specific Analytics
    • Sports Betting and Business Analytics
  • Cited References
• Glossary
• Index

pokaż cały spis treści

ukryj recenzje