Text Analysis

Text Analysis

Dan Petty and Jack Grobe

We looked at the words “slave(s)”, master, man, house, field, and women. we chose these words in an attempt to view correlations between gender and space. We struggled quite a bit with this assignment but there is value in the struggle. While there was very little we could determine from our analysis of the word groupings we realize the value of each tool. The first tool that we used is the trends line which shows the frequency of words in each document in the corpus. The biggest change that we see when looking at the trends line for our words is the drastic shift between the use of master and slave from 1760’s to the 1860’s. It is hard pick a reason for this shift but the trends line makes the difference quite clear.

Knots uses time to represent length of a text and terms as lines that extend out. Each time a term is mentioned the line will bend creating what looks like knotted lines across the screen. From an analyzation standpoint, knots are a useful representation as it visually depicts frequency not simply as a static number but as distance between uses. For example, a knot of the 1862 Jackson work, we see two very different types of lines. The first being the long arcs with slight bends of both house and field (the blue and dark blue lines) which points to their placed as settings in the narrative as the bends of field line up with the straight lines on house. The other type of line is that of slave(green) and in some cases master(yellow). The kinks and bends of the slave line points to the term as the central theme of the document. Also, unlike field and house, there is no inverse relationship between slave and master.

I had problems with the site when I tried to input my own search terms to be analysed. I tried to search terms by entering them in the search bar, which took a while to load. When they did load and I tried to add them into the search, the Voyant system would throw an error message. Therefore, I had to use the terms already in the correlation analyser already as seen above. I sorted the terms by correlation, to see what terms are used in inverse with “woman”. Terms that approach -1 have an inverse correlation, 1 have a standard correlation, and 0 are correlationally meaningless. The terms that had the most significant correlation with the term “woman” were “soon”, “mr”, and “away”. I had expected more gendered words or work terms to be the most inverse with woman, instead of the seemingly random words “soon” and “away”.

Comments Off on Text Analysis