While I did find this lab to be interesting, I did face several challenges while completing it. The first challenge I faced while trying to complete this lab was just simply wrapping my head around the concept of topic modeling and how Mallet works. After finally feeling like I had a better understanding of this program and how it is used to supplement literary analysis, I was able to then start using Mallet through my terminal application on my Mac. The second challenge I faced was with the process of actually inputting data to the terminal. It was hard for me to remember all of the steps we originally took in class in order to begin to use Mallet to collect data. For example, I simply began the lab by entering the first command in terminal without actually choosing Mallet as the preferred directory. Once I was able to get past this hurdle, I then moved on to entering the correct commands. Other than these two hiccups, properly using Mallet seems to heavily rely on making sure all commands are entered correctly. It’s really interesting that even just one missing space or dash could cause a whole command to be invalid.
Topic modeling as a practice within digital literary studies is something that is very intriguing to me. I think it helps answer and investigate certain questions that often arise when studying literature. If topic modeling is about viewing overarching ideas, themes, or words/phrases within a group of documents or texts, it really helps answer questions without having to put in the time or work of actually reading each of those texts. Topic modeling helps answer questions like these: What are some overarching ideas in my group of texts? How often do these words appear in my group of texts? In my individual texts? What percentages of my texts are made up of each specific topic? While these specific questions don’t cover all of the questions that topic modeling can answer and explain to researchers, they are the ones that have stood out the most to me in my experience with Mallet and other topic modeling tools.
Some of the topics from my fantasy corpus seem to be a lot more coherent than other topics. For example, topic 18 clearly makes sense as it stands in the fantasy corpus: wizard, magic, Dorothy, etc. Also, topic 12 lists: man, king, young, princess, etc. These topics appear to be the most coherent because they are ideas and themes typically associated with fantasy works of literature. On the other hand, other topics don’t seem to make as much sense. Topic 3 doesn’t seem to be very coherent. It lists: direction, distant, door, conversation, return, etc. as its ideas and words to be seen throughout the texts. It seems as though coherency really boils down to what our preconceived notions are about a type of literary genre or other aspect of a literary corpus. If we do not know much about a genre, however, it could be much more difficult to pinpoint inconsistency.