ggdag

ggdag is an R package designed for the visualization of directed acyclic graphs (DAGs). It leverages the ggplot2 framework to create aesthetically pleasing and informative DAG plots, commonly used in causal inference and statistical modeling. The package allows users to specify nodes and edges of a DAG, and then renders them using ggplot2's grammar of graphics. This enables customization of node shapes, colors, and labels, as well as edge styles and arrowheads.

Key features of ggdag include the ability to represent causal relationships, identify confounding variables, and illustrate the structure of statistical models. It can automatically detect and draw common DAG patterns, such as colliders and mediators, which are crucial for understanding potential biases in observational data. The package also integrates with other R packages used in causal inference, such as dagitty and bnlearn, facilitating a seamless workflow. ggdag aims to make the complex task of visualizing and interpreting DAGs more accessible to researchers and data scientists. Its foundation in ggplot2 means users familiar with that ecosystem can quickly adapt to its syntax and capabilities for creating publication-ready DAG diagrams.