This project was created for the midterm of DGAH 110: Hacking the Humanities. In this short project, I used Voyant Tools to conduct textual analysis on the journals of Meriwether Lewis and William Clark, as they traveled westward on their expedition of North America. I aimed to identify certain relevant trends in the text corpus. Specifically, I wanted to see how the text data that could be extracted from the journals lines up with our existing knowledge of the journey.
All text data used for the parts of this project is from the Project Gutenberg eBook of The Journals of Lewis and Clark, 1804-1806. Before doing analysis of the text, I had to do a small amount of data cleaning. This consisted of removing the appended and prepended text that Project Gutenberg adds, so that Voyant would not include these unnecessary parts of the text in the corpus.
I began this text analysis project by uploading my corpus into Voyant Tools, an online platform for conducting text analysis. I chose to use this tool because of its relative ease of use, as well as the fact that I gained some experience earlier in the course using it. After uploading my corpus, I played around with the suite of tools offered by Voyant, until settling on about three that I thought effectively conveyed insightful information about the source text. I wanted to ensure that the visualizations both conveyed information about the text, but were not overly full of fancy effects or too packed with information. The visualizations that I ended up settling on were Trends, TermsBerry, and WordTree. While these tools all show similar metrics, they pull apart the information in different ways that lend themselves to different kinds of analysis.
When presenting my findings, I chose to do so making use of the Hugo static site generator. I felt like creating a WordPress site was overkill for a project of this simplicity, and having used Hugo in the past, I thought this was the perfect use case. Hugo takes markdown files and then generates HTML, meaning that adding content to my site is extremely easy. Additionally, I chose to embed my Voyant Tools visualizations in this site directly rather than adding images, which increases the interactivity and viewer engagement.
While the results of applying text analysis to the Journals of Lewis and Clark aren’t particularly groundbreaking, they support trends that we already know about the text, and aid someone unfamiliar with the source material in understanding it. Using frequency analysis, for example, we can see that the word “river” is the single most frequent word throughout the corpus. This shows just how important rivers were to Lewis and Clark, as they relied upon them for transportation. Furthermore, using collocation analysis, we can see three words that very commonly co-occur are “deer,” “elk,” and “killed.” This indicates that the travelers almost solely relied upon deer and elk for sources of meet. This project, more generally, shows that this process of distant reading is highly effective for pulling out trends from text corpora. While a purely statistical technique of analysis might be effective as well, the Journal of Lewis and Clark is an extremely messy dataset. This means that, to some extent, analyzing it is less of a science and more of an art - it takes a human mind to interpret and make sense of.