Hopefully I did this right.
I got rid of the table of contents mainly because it was not needed for this project. I also removed all the headings and page numbers. Since I obtained this source from JSTOR I kept removing a very long heading that was adding a lot of useless numbers and words unrelated to the topic.
The main ones I chose were Burnside, operation, army, soldiers, navy, division, and war. These words are important to the main point of my project looking into Burnside’s military career.
Bubble line was really useful because it gave me a more accurate location for certain words throughout the document. The reader tool also helps for the same reason, highlighting the words that I am specifically looking for. Cirrus was pretty neat mainly because it brought to my attention the most frequent terms.
I was expecting more correlations but unfortunately they only presented singular completely unrelated correlations that didn’t help me at all.
The only impact it really gave me was allowing me to sift through the text to find the parts of the history relevant to my topic.
Reading texts in future I will definitely be looking for what words seem to be most prevalent within the text and one thing I may do better than voyant is try to notice better correlations.
Well as discussed above I think their correlation tab isnt the most useful. Not sure how it can be misused but it definitely could confuse someone with the amount of tools and information it tries to present.
It reveals patterns in text which can be useful when looking for specific words that relate to the topic. The limitations of text mining are mainly that the results can be skewed if youre trying to look for something specific. For example in my project Im researching burnside but I dont need text mining to find the portion of the text that mentions burnsides name but nothing relevant to his career.
Long, David E. “Burnside When He Was Brilliant: Ambrose Burnside and Union Combined Operations in Pamlico Sound.” In Union Combined Operations in the Civil War, edited by Craig L. Symonds, 10–22. Fordham University Press, 2010. https://doi.org/10.2307/j.ctt13wzx8w.7.
