Εxcited to share TextAssociations.jl, a new Julia package designed for studying how words connect and co-occur in texts and corpora. It brings together many well-known and lesser-known association measures, along with tools for collocation networks, temporal analysis, and large-corpus processing, all in one unified, data-driven framework.
This package represents a first attempt to introduce Julia to the humanities and social sciences, with a focus on researchers working in linguistics, digital humanities, and language-centered social science research. It’s a starting point and if the idea resonates with you, I’d really appreciate it if you could give the GitHub repo a star.
In these fields, we are often interested not only in meaning but also in pattern and context — how words cluster within a discourse or register, across time periods, or within specific language varieties such as dialects and learner corpora. While embeddings and RAG-based systems operate within pretrained semantic spaces, association measures remain indispensable for interpretable, corpus-based insights: they quantify how often words co-occur beyond what chance would predict, revealing what is salient within a given dataset, discourse, or community.
The package is currently in an early release stage (not yet registered in the Julia package registry), and there are still some rough edges to smooth out. Feedback, contributions, and testing would be greatly appreciated.
GitHub repository: GitHub - atantos/TextAssociations.jl
Documentation: Home · TextAssociations.jl
