I looked at jsLDA and their goal is to model data in text. It takes documents (not documents in the way we know them, but rather, chunks of texts) and uses various stop words to sort themes. For example one that the website talks about is studying how certain themes came up in the state of the union address and classifies them by category.
In my opinion the project was very cool and it even had a user-end interaction option. However, the website lacked substance. The project was there but the presentation lacked anything of any real excitement value. Additionally, the site only has one page. It offers little information, poorly explains the purpose of the project, and presents it’s user interface in such a way that it seems almost disinteresting.