I chose to analyze and reverse engineer an experiment that utilizes word vectors to study the ever changing standard of beauty in Vogue in a project titled Robots Reading Vogue.
Sources: Vogue Archive launched by ProQuest and Condé Nast & Yale University Library via ProQuest
Processes: A temporal word embedding algorithm, called Word2Vec, was used to capture highly used words and phrases from the many different authors and editions of Vogue. This was in the effort to understand how beauty was portrayed in the magazine by text throughout different decades. This algorithm used contemporary linguistic theories to find how word are associated with each other by their relative position to each other. The (cosine) distance between pairs of words was measured and ranked to reveal semantic neighbors.
Presentation: The data for each decade was organized into a dendrogram, which is a diagram that shows the hierarchical relationship between objects. A 2D map with dimensionality reduction was also used to present the data as a whole and let users explore and interact with the findings. An example of a dendrogram can be seen below.
A new question I have after analyzing this project is would this same technique of finding trends through connected words work with less expansive works, such as a singular book or article?
Who made the website? What are their relationships to the institution? Sydney Bowen made the project which is a part of a larger collection of experiments, coined Robots Reading Vogue, in the Digital Humanities department at Yale University Library. Robots Reading Vogue mines data in fashion to explore topics in various areas of study, ranging from computer science to gender studies. Bowen was a student at Yale University who was a part of the 2021 class.
Which academic fields (i.e history) do you see the project in conversation with? This project could be related to history, fashion, linguistics, and potentially even sociology.
2 thoughts on “Analyzing DH Projects – Lab 2”
Leave a Reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
What an interesting project! I wonder what this project would look like compared to other beauty or fashion magazines – what would be the differences? I also explored this website and found the 2D map you mentioned to be such an interesting and cool way to explore this type of information, and I am curious about how it could be applied to other projects in the future.
This is such a cool project! I think you did a great job explaining how this website works, as well as how the computers collect and show the user data. I had never heard of a dendrogram, so that was very interesting, and you did a great job explaining that! I wonder if this could be used further in gender studies to understand the difference in language used to describe men and women!