For my Voyant textual analysis, I chose the book The Secret Garden, by Frances Hodgson Burnett (available at Project Gutenberg). I chose it because when I was younger, my mom strongly recommended that I read it. When I read it, it must not have been impressionable because I don’t remember much about it. But, for this reason, I thought it would be a good choice for this blog assignment because I’m curious about how much I can recall about this story by doing a textual analysis of the book.
Tool #1: Trends
The first approach I took was using the ‘Trends’ tool, which shows the relative frequency of a particular word over the course of the text. I like this tool because it could indicate the plot at certain times or even the tone.
I think this visual is fascinating. Here I compared the word ‘said’ (this is the most commonly used word in the text, 832 usages) to the words ‘secret’ (77 usages) and ‘garden’ (232 usages). I expected that the words used in the title, especially ‘secret,’ would be used far more frequently than they are actually used. Another interesting point I’ve extracted from this visual is that there seems to be more dialogue in the second quadrant of the book, and very little dialogue at the beginning and end of the book I wonder why this is. So far I don’t think I can recall much about the story from my textual analysis.
Tool #2: Links
The second tool I tried out is called Links and it calculates the words most often associated, or used close to, the word of choice. I like this tool because it shows the various contexts in which this word was used.
I like this visualization because there seems to be an endless amount of connections to be made. By clicking and dragging the nodes, the whole graph can shift, showing a whole new perspective on the same relationships. Above, I chose the five most frequent nouns that could indicate place (an attempt to recall the story). Again, I didn’t recall much about the plot but I think its interesting that many of the orange words are not connected to more than one blue word, only ‘Mary’ and ‘said.’
I wonder how powerful tools like Voyant are. For example, if I had an unordered list of words and frequencies, would a software be able to calculate the most probable storyline? What if it had the data like in the Trends tool? I wonder how much more powerful this software can and will be. For now, it seems like a pretty cool tool for people like myself to use for fun and curiosity.
2 thoughts on “The Secret Textual Analyzer”
Leave a Reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
I thought your analysis of the dialogue was clever. I am interested to see if it is actually true that the lack of the word “said” is directly linked to dialogue. I could imagine that there might be other verb tenses or shifts in the tone of the book that could reduce the word “said” without removing dialogue. Regardless, it is a very interesting spike.
I too cannot recall any information about the plot of the Secret Garden, despite my mother’s best efforts. I thought your comments were interesting about a text analysis software potentially being able to create a story from a random dataset of words. I never considered that kind of application – I only had considered using it to reveal author biases.