We propose to reconstruct causal graphical models from textual data via a new Python package, WordGraph. This package facilitates the exploration of large corpora of documents through interactive visualizations in the form of graphical word models, which is available via a GitHub repository.