r-causal/ggdag

:arrow_lower_left: :arrow_lower_right: An R package for working with causal directed acyclic graphs (DAGs)

What it solves

It provides a way to tidy, analyze, and visualize causal directed acyclic graphs (DAGs) within the R ecosystem, specifically making the dagitty package compatible with the tidyverse and ggplot2 workflows.

How it works

The package leverages dagitty to create and analyze structural causal models. It allows users to define DAGs using either a specialized syntax or a more R-like formula syntax via dagify(). It then converts these DAGs into "tidy" data frames that can be plotted using ggplot2 and ggraph through a set of dedicated geoms and helper functions.

Who it’s for

Researchers and data scientists using R who need to map out causal relationships and identify adjustment sets for causal inference.

Highlights

  • Uses dagitty for the underlying causal analysis.
  • Integrates with ggplot2 for flexible, high-quality visualization.
  • Supports the creation of DAGs via a tidy-style formula syntax.
  • Includes built-in functions for visualizing common bias structures like butterfly bias and confounder triangles.

Related

  • Project
    pgmpy/pgmpy

    A Python library for causal and probabilistic reasoning using graphical models, providing tools for structure learning, parameter estimation, and causal inference.

  • Project
    juliasilge/tidytext

    An R package for text mining that converts text into tidy, one-token-per-row data frames for easier analysis with standard data manipulation tools.

  • Project
    ChunelFeng/CGraph

    A cross-platform C++ DAG framework that allows developers to build complex task pipelines with dependency management, parallel execution, and Python API support.

  • Project
    stan-dev/bayesplot

    An R package providing a library of plotting functions for visualizing posterior draws, MCMC diagnostics, and predictive checks for Bayesian models.

  • Project
    cmu-phil/tetrad

    Tetrad is a software suite for causal discovery and analysis, offering a GUI application and integrations for Python and R to identify causal relationships in data.