Emilia Rodriguez
Introduction
This project was created for my Digital Humanities midterm that asked us to clean a dataset, perform some type of analysis or digitization, and embed our results on our webpage. I chose to look at Zitkala-Ša’s “American Indian Stories” as I’ve taken a couple classes related to Indigenous tribes and thought this would be an interesting continuation of these studies. Since I chose a text, I performed text analysis on these stories using Voyant Tools and created several visualizations depicting the most frequently used words, as well as looking at different frequencies of specific types of terms. The results and description of this project are below.
Most Frequently Used Words
Most Frequently Used and Collocated Words
Frequency of Words Related to Race and Nationality
Frequency of Words Related to Gender and Roles
Sources
These visualizations were created using Voyant Tools on Zitkala-Ša’s “American Indian Stories”. Before uploading the dataset, I first separated the initial .txt file containing the complete text into separate .txt files for each section listed in the contents. Otherwise, Voyant Tools will automatically separate the dataset into sections based on equal word counts rather than by how the stories are logically separated in the text. Further, I had to remove the licensing agreement at the end as it was longer than most of the stories and significantly altered the results.
Processes
Using Voyant Tools, I then added more words to the default stop words list to further refine the visualizations to represent the most meaningful words based on what words I knew would be consistently repeated in the list of authors and contributors. For my visualizations, I chose to create a Cirrus, TermsBerry, and a couple Trends graphs. For the Cirrus and TermsBerry, I further refined the stop words list to filter out any words that appeared that weren’t meaningful, such as “said”. For the Trends graphs, I looked at the most commonly used words and chose to compare the most frequently used words related to race and nationality: “indian”, “american”, and “white”. Then, I chose to compare the most frequently used words related to gender and roles: “mother”, “woman”, and “man”.
Presentation
For the website, I chose to use bright colors to match with the colorful palette of Voyant visualizations. First, I chose to embed the two more general visualizations that represented the text as a whole before the trends throughout different sections in the book. Then, I chose to embed the two trend visualizations in a row next to each other as they were examining similar categories of terms relating to social disparities. Finally, I included labels for all of the visualizations to make each visualization more clear without additional context.
Significance
Digital text analysis can help gain new insights about the importance of terms to different communities. For example, I chose the categories for the Trends graphs after noticing these words in the Cirrus as I thought these categories of terms would be relevant to Indigenous stories. Firstly, since colonization and the establishment of America has caused innumerable and significant loss to Indigenous peoples, comparing the terms related to the nation, the white people that colonized it, and the term representing Indigenous peoples could provide insight into how colonization has impacted sense of identity within Indigenous peoples. Further, colonial and tribal societies think differently about gender, with colonial societies typically being patrilineal while Indigenous communities are typically organized matrilineally. Thus, the comparison of the frequency of usage of the different categories related to gender can provide valuable insight into this different organanization of society. Even the fact the that the word “father” was not used frequently enough for it’s inclusion is an interesting insight. As opposed to data science in general, with the context Arts & Humanities studies provide, this approach is also intentionally more accessible to those without data science knowledge, which includes some Indigenous peoples as they have historically been restricted or discouraged from pursuing higher education. When dealing with data that involves human stories, especially when the communities represented have faced oppression and violence, it is more important than ever to use humanities informed methods that consider people beyond their statistical significance. Digital humanities, if used correctly, can be a tool of restorative justice for these communities in it’s accessibility and customization abilities. Overall, this project illuminates how Digital Arts and Humanities can be an important tool for modern ethical engagement with harmful histories.