This blog post computationally explores the latent topics present in a selection of public case summaries from the Refugee Appeals Board. It showcases how natural language processing can be used for initial exploration of text data as well as some of the considerations that must go into selecting which model to employ. I present a few findings and how one can go about using topics to discover cases, explore topological distribution for a specific group of applicants, and how topics are associated with decisions of the Board.