Using Voyant as an Entry Point for Analyzing Trends in Lit Mags
For creative writers, one thing that can be tricky is trying to get a sense of a particular journal’s aesthetic, and figuring out whether to submit. As I considered the way that Voyant presents and organizes information, I wondered if it would be possible to map trends or patterns of language within a specific journal. As an experiment, I undertook this sort of analysis on a very small scale. I selected the first ten poems in issue 37 of Rattle, and entered them into Voyant for analysis. My objective was to get a sense of where the language of these poems overlapped, and use this information as a starting point for considering potential trends or patterns in themes and content of the issue in question. I stress that I consider this only a starting point, as I would need to read and analyze the entire journal in order to draw meaningful conclusions about these trends or patterns. While my project only offers information on a single issue of the publication, it would also be feasible to expand the results by analyzing multiple issues of Rattle spanning several years. Instead of providing a starting point for mapping trends and patterns within a specific issue, this would map these elements within the context of the journal, as a whole, during a specific time period.
The first thing I noticed, after inputting my text, was that the second most frequent word was man, with 5 uses in ten poems. (The most frequent was “like,”which occurred 8 times, over six different poems.)
After noticing this, I investigated to see which poems “man” occurred in. Perhaps coincidentally, all these poems were written by men, although several poems by women were included in the sample I chose.
While this information is not enough to logically draw the conclusion that male authors in this issue are more frequently writing about gendered topics than their female counterparts, it makes me want to investigate further by reading the poems in question. This is an example of how these visualization tools, through offering a “big picture” overview, can help narrow a reading focus. I might now read this issue of Rattle with special attention to how male and female writers address gender in their poems. Incidentally, I did look at all uses of “man” in the corpus reader view of the text, and can confirm that all instances were in reference to a male person, although further analysis would be required to determine the significance of this. While it was necessary to read the text in order to verify this, the data in Voyant did point me in the right direction.
Another thing I noticed was that the frequency counts for words were quite low, which is probably because I was working with such a small sample size. However, I did notice that the total word count was 1,666, with 850 of them listed as unique. This data raised some interesting questions about word frequency within poetry, as opposed to prose. I’m curious as to whether the variety tends to be greater in poetry than in other texts. While the data I got from Voyant isn’t enough to support a conclusion, it did give me food for thought that I wouldn’t have otherwise had. It pointed me in a direction that could be considered for a future research project. Again, the macro focus seems to play a key role in how data generated by Voyant can serve as a sort of compass.
As a recap, my initial purpose was to determine whether data produced by Voyant might be useful in the context of identifying content or thematic trends within an issue of a poetry publication and, in a larger context, the publication as a whole (within a defined time frame.) I do think that my analysis of the data supports the idea that Voyant can be useful in this regard, although it’s clearly necessary to employ other analysis techniques as supplements. As a sidebar, my analysis of the data also gave me an interesting idea for doing a comparison between word frequencies of poetry and prose texts, which was an added benefit of using this visualization tool.