Causality in Electricity Markets

Overview

Causality in Electricity Markets is a comprehensive guide to understanding and applying causal inference techniques in electricity markets. The book provides a structured introduction to causal reasoning, statistical modeling, and machine learning methods for uncovering cause-and-effect relationships in energy systems.

Why This Book?

In modern electricity markets, predictive models alone are often insufficient. Understanding causal mechanisms is essential for:

• Identifying true drivers of market behaviors.
• Developing effective policies and interventions.
• Improving forecasting models with causal structure.

This book bridges the gap between causal theory and practical applications, offering real-world case studies in electricity market analysis.

Table of Contents

The book is divided into five key sections:

1. Introduction to Causal Inference

   • Causal vs. predictive models
   • Directed Acyclic Graphs (DAGs)
   • Structural Causal Models (SCM)

2. Causal Discovery

   • Learning causal structures from data
   • Semi-parametric algorithms: LiNGAM, VAR-LiNGAM, ANMs

3. Causal Inference Methods

   • Instrumental Variables (IV)
   • Propensity Scores and Matching
   • Double Machine Learning (DML)

4. Interpretability & Market Applications

   • Shapley values & Partial Dependency Plots
   • Impulse response functions
   • Case studies in electricity markets

5. Experimental Design & Data Collection

   • A/B testing
   • Bandit algorithms
   • Active learning

Getting Started

Requirements

To reproduce examples in this book, install the following Python libraries:

pip install numpy pandas statsmodels scikit-learn dowhy matplotlib networkx

Running the Notebooks

Jupyter notebooks are provided for hands-on learning:

git clone https://github.com/your-repo/causality-in-electricity-markets.git
cd causality-in-electricity-markets
jupyter notebook

Authors

Davide Cacciarelli
Research Associate, Imperial College London
Expert in causal inference, machine learning, and energy markets
Personal Website | Email

Pierre Pinson
Chair of Data-centric Design Engineering, Imperial College London
Editor-in-Chief, International Journal of Forecasting
Personal Website

Citation

If you use this book in your research, please cite:

@book{cacciarelli2024causality,
  author    = {Davide Cacciarelli, Pierre Pinson},
  title     = {Causality in Electricity Markets},
  year      = {2024}
}

Contact

For inquiries, please reach out to d.cacciarelli@imperial.ac.uk.