For this week’s blog assignment, I downloaded the text for Oliver Twist by Charles Dickens from Project Gutenberg, a free online library with access to over 70,000 ebooks.
voyant
For my first experiment with textual analysis, I explored the many different options in Voyant Tools. I liked the visualization tools such as Cirrus and TermsBerry as you could easily identify the key words that were most frequently used in the text.
I chose to explore the TermsBerry tool more closely as it provides information on not just the frequency of the key word, but also where and with what other words the key word is grouped in. To be specific, if you hover over one of the words you want to explore in the below visualization, it will not only tell you the number of times it appears in the text, but it will also show you which other words are used close to the word of interest, with darker pink signaling the greatest proximal relationship.
gemini
The second experiment was done using Google Gemini, an AI tool. I followed the tutorial by Anastasia Salter which guided me through how to use Gemini to feed it the text, filter stop words, and create a meaningful visualization in the form of a world cloud. One small issue I ran into was that Gemini took a long time processing the text file and each instruction I gave it, and I had to reload the page and try again several times before it would give me an output.
concerns about ai
While AI is a useful tool for those who are self-taught, there are some concerns that should be taken into account. A main concern I have is the lack of critical thinking AI can lead to. By feeding it the necessary text and giving it some instructions, it will produce an output. However, AI can oftentimes be incorrect or the instructions you give it may not be detailed enough or well specified, resulting in outputs that are just flat out wrong or that are different to what you actually want it to produce. If we were to rely on the results, this could lead to incorrect conclusions of our analysis which is problematic. We should always remain cautious and cross check whenever we can to ensure valid results.