Blog Post Week 2

• Introduces and links to the project you explored.
• Contains at least one image of the project that links directly to it
  • If you don’t know how, follow these instructions to take a screenshot on a Mac/on a PC.).
• Breaks down the black box of your digital project by indentifying its
  • Sources (assets)
  • Processes (services)
  • Presentation (display)
• Identifies a new question you have that arose from breaking the project down
• Answers at least two of the in-class discussion questions above

The project that I chose to explore was Robots Reading Vogue, which explores data mining in fashion through archived issues of Vogue magazine. Specifically, the section that I looked at within this project was using temporal word embeddings to reveal the shifting notion of beauty in Vogue. This project looked at the words that were associated with the word ‘beauty’ in issues of Vogue magazines, and how this changed over different periods of time. It also provided examples of how beauty was written about within the context of the articles.

Source: The data source for this project was Yale University Library, which had full access to the Vogue Archive, a digital collection of over 400,000 pages created in 2011 of all the Vogue magazines that have been published since 1892.

Processes: The author of this project used the word embedding algorithm Word2Vec, which employs a neural network to find the semantic similarity between words.

Presentation: The data is presented on a website with images of word trees that show the semantic partners of the word ‘beauty’ in different periods of time. These images are supplemented by text describing the words seen in the tree and give context and example quotes. The website also has paragraphs describing the larger insights from the data before and after each word tree.

A question that I have that arises from this project is how the Word2Vec works in detail, and how it can be used to make interesting observations about very large data sets. Based on my background experience with neural networks, it seems like these kinds of analysis tools could be leveraged to analyze big data sets without intensive human work, and I want to explore more on how that works.

This project has lots of commentary with the data, and it makes the argument that beauty, and the words that are associated with beauty, have changed radically over time. Specifically, the project argues that the idea of beauty shifts from a paradigm that embodies one’s intellectual capacities and mannerisms to one that is overwhelmed by commercialization and the advent of modern-day cosmetics and beauty products as represented by the stereotypical white woman.

The target audience for this project is quite large, as the idea of beauty in society is something that is important for the general population to be aware of. Especially in a time where social media can put a lot of social pressure on women in society, this project can also help to show how beauty is very subjective and changes very quickly, which can help assure those who struggle with how they fit into the mainstream definition of beauty that there is no one right answer to what is beautiful. This data specifically would also be of use to any scholar looking at the history of popular opinion and beauty standards over time and gives a very clear window into how it changes and progresses.