unravelsports

Getting Started

  • Installation
    • Basic Installation
    • Requirements
    • Optional Dependencies
      • For Graph Neural Networks with PyTorch
      • For Graph Neural Networks with Spektral (Python 3.11 only)
    • Development Installation
    • Getting Help
  • Quick Start
    • Soccer: Train a GNN
      • Step 1: Load Data with Kloppy
      • Step 2: Convert to Polars DataFrame
      • Step 3: Convert to Graphs for GNN Training
      • Step 4: Train a Graph Neural Network
    • American Football: BigDataBowl Data
    • Soccer Analytics Models
      • Pressing Intensity
      • Formation Detection (EFPI)
    • Next Steps
  • Core Concepts
    • Data Flow
    • Tracking Data
      • What is Tracking Data?
      • Supported Data Providers
    • Polars DataFrames
      • Why Polars?
      • DataFrame Structure
    • Graph Neural Networks
      • What are Graphs?
      • Why Use GNNs for Sports Data?
      • Graph Conversion Settings
    • Labels and Graph IDs
      • Labels
      • Graph IDs
    • Soccer Analytics Models
      • Pressing Intensity
      • EFPI (Elastic Formation and Position Identification)

Tutorials

  • Soccer Graph Neural Networks
    • Interactive Notebooks
      • Quick Start Guides
      • In-Depth Tutorials
      • Graph Configuration FAQ
    • Key Concepts
      • Data Preparation
      • Graph Conversion
      • Converting to Graph Format
      • Model Training
      • Making Predictions
  • Pressing Intensity
    • Interactive Notebook
    • Basic Usage
      • Step 1: Load Tracking Data
      • Step 2: Initialize Model
      • Step 3: Compute Pressing Intensity
    • Parameters Explained
      • Time Window
      • Method
      • Ball Method
      • Orientation
      • Speed Threshold
    • Output Format
    • Visualization
  • Formation Detection (EFPI)
    • Basic Usage
      • Step 1: Load Tracking Data
      • Step 2: Initialize EFPI Model
      • Step 3: Detect Formations
    • Parameters Explained
      • Formations
      • Time Granularity (every)
      • Substitutions
      • Change Threshold
      • Change After Possession
    • Output Format
  • American Football (NFL)
    • Interactive Notebook
    • Data Format
      • Big Data Bowl Data
    • Basic Usage
      • Step 1: Load Data
      • Step 2: Add Labels and Graph IDs
      • Step 3: Convert to Graphs
      • Step 4: Train a Model
    • Data Availability

API Reference

  • Classifiers
    • PyTorch Geometric
    • Spektral
      • CrystalGraphClassifier
        • CrystalGraphClassifier.__init__()
        • CrystalGraphClassifier.call()
    • Usage Examples
      • PyTorch Geometric
      • Spektral
  • Soccer
    • Dataset
      • KloppyPolarsDataset
        • KloppyPolarsDataset.data
        • KloppyPolarsDataset.settings
        • KloppyPolarsDataset.home_players
        • KloppyPolarsDataset.away_players
        • KloppyPolarsDataset.kloppy_dataset
        • KloppyPolarsDataset.__init__()
        • KloppyPolarsDataset.convert_orientation_to_ball_owning()
        • KloppyPolarsDataset.load()
        • KloppyPolarsDataset.add_dummy_labels()
        • KloppyPolarsDataset.add_graph_ids()
        • KloppyPolarsDataset.get_player_by_id()
        • KloppyPolarsDataset.get_team_id_by_player_id()
        • KloppyPolarsDataset.sample()
    • Graphs
      • SoccerGraphConverter
        • SoccerGraphConverter.settings
        • SoccerGraphConverter.n_node_features
        • SoccerGraphConverter.n_edge_features
        • SoccerGraphConverter.n_graph_features
        • SoccerGraphConverter.dataset
        • SoccerGraphConverter.chunk_size
        • SoccerGraphConverter.non_potential_receiver_node_value
        • SoccerGraphConverter.edge_feature_funcs
        • SoccerGraphConverter.node_feature_funcs
        • SoccerGraphConverter.global_feature_cols
        • SoccerGraphConverter.global_feature_type
        • SoccerGraphConverter.additional_feature_cols
        • SoccerGraphConverter.default_node_feature_funcs
        • SoccerGraphConverter.default_edge_feature_funcs
        • SoccerGraphConverter.get_players_by_team_id()
        • SoccerGraphConverter.get_player_by_id()
        • SoccerGraphConverter.plot()
        • SoccerGraphConverter.__init__()
        • SoccerGraphConverter.settings
    • Models
      • PressingIntensity
        • PressingIntensity.output
        • PressingIntensity.dataset
        • PressingIntensity.chunk_size
        • PressingIntensity.fit()
        • PressingIntensity.__init__()
      • EFPI
        • EFPI.output
        • EFPI.segments
        • EFPI.return_dtypes
        • EFPI.__init__()
        • EFPI.fit()
        • EFPI.dataset
      • Formation Detection (EFPI)
  • American Football
    • Dataset
      • BigDataBowlDataset
        • BigDataBowlDataset.data
        • BigDataBowlDataset.settings
        • BigDataBowlDataset.__init__()
        • BigDataBowlDataset.load()
        • BigDataBowlDataset.add_dummy_labels()
        • BigDataBowlDataset.add_graph_ids()
    • Graphs
      • AmericanFootballGraphConverter
        • AmericanFootballGraphConverter.dataset
        • AmericanFootballGraphConverter.settings
        • AmericanFootballGraphConverter.label_column
        • AmericanFootballGraphConverter.graph_id_column
        • AmericanFootballGraphConverter.__init__()
        • AmericanFootballGraphConverter.return_dtypes
  • Utils
    • Objects
      • GraphDataset
    • Features
      • graph_feature()
      • x_normed()
      • y_normed()
      • speeds_normed()
      • velocity_components_2d_normed()
    • Helpers
      • dummy_labels()
      • add_dummy_label_column()
      • dummy_graph_ids()
      • add_graph_id_column()
      • Adding Graph IDs
    • unravel.utils.dummy_labels
      • dummy_labels()
    • unravel.utils.dummy_graph_ids
      • dummy_graph_ids()
    • unravel.utils.add_dummy_label_column
      • add_dummy_label_column()
    • unravel.utils.add_graph_id_column
      • add_graph_id_column()

Additional Information

  • Citations
    • The Package
    • Graph Neural Networks
    • Pressing Intensity
    • Formation Detection (EFPI)
  • License
    • Summary
    • Full License Text
    • Key Requirements
    • Using in Closed-Source Projects
    • Questions?
unravelsports
  • Overview: module code

All modules for which code is available

  • unravel.american_football.dataset.dataset
  • unravel.american_football.graphs.graph_converter
  • unravel.classifiers.crystal_graph
  • unravel.classifiers.crystal_graph_pyg
  • unravel.soccer.dataset.kloppy_polars
  • unravel.soccer.graphs.graph_converter
  • unravel.soccer.models.formations.efpi
  • unravel.soccer.models.pressing_intensity
  • unravel.utils.features.builtin
  • unravel.utils.features.utils
  • unravel.utils.objects.graph_dataset
  • unravel.utils.utils

© Copyright 2024, Joris Bekkers.

Built with Sphinx using a theme provided by Read the Docs.