SocSemVar

Analysing semantic change and socio-semantic variation on Reddit.

This repository contains the code for the paper Semantic change and socio-semantic variation. The case of Covid-related neologisms on Reddit, published in the journal Linguistics Vanguard.

You can clone the repository and install the code as a Python package named socsemvar by running pip install . within the cloned directory. This will automatically install all dependencies. As always, it is recommended to install this package in a virtual environment (e.g. using conda).

The Reddit data used for this paper are too big to make them available here. Therefore, the code in this repository uses a smaller random sample of the original data. The full datasets of Reddit comments and the models trained from these comments can be requested via email and allow to reproduce our results.

This repository provides the code used to process the Reddit comments, train the models, and produce the results presented in our paper. The code was written and documented using the literate programming framework nbdev and the documentation is available here:

https://wuqui.github.io/socsemvar/