1.
Erasing non-informative words: “ad” and “hoc”.
2.
3.
Top ten most frequent words per topic for the CTM with 13 topics.
Overview of 2021 master programs: Distribution by institution and corpus of student body by enrollment numbers, categorized into general (G) or specific (S) programs.
The posterior distribution of the 7 topics across the 4 master programs offered at Utrecht (UU, top) and Groningen University (RUG, bottom).
Posterior classification over the 7 topics of the courses offered in the 4 master programs at Utrecht University (UU) and Groningen University (RUG).
Plot of the coherence scores per number of K topics for the CTM.
Seven bar plots of the posteriors per university for every manually labelled topic of the CTM outcomes.
The posterior distribution per program type (general vs subject specific) of the 7 topics of the CTM model.
Number of program and course descriptions per university included in the analysis.
Manual changes made to the texts as part of data pre-processing.
