Germain Gauthier

Relatio is a Python package to extract simple narrative statements from large text corpora (e.g., “Taxes ruin the economy.”, “Democrats stole the election.”, “People are losing their jobs.”).

For a given corpus, we map semantic relationships between K latent entities. The number of latent entities is chosen by the user (we offer quantitative metrics for a data-driven choice). Our package then allows users to plot narratives as directed multigraphs (some examples). The method is described in greater detail in the original paper.

Check out the PyPi version and the latest development version for your own applications.