Word Cloud

This word cloud made using Voyant takes all the artwork titles from the spreadsheet and allows me to see how frequently certain words were used to name each piece.

After seeing this word cloud, I wondered how different it would look if I removed prepositions, conjunctions, and articles, essentially leaving only the more unique verbs, adjectives, and nouns.

Now we can see more about the subjects of the paintings based on their titles. What if I got rid of terms that indicate the medium of the artworks, like "drawings" or "portrait?"

Here we can see even clearer what the subjects of the artworks are. The last thing I want to try is to translate all of the artworks into English. For example, I see that both "woman" and "femme" appear here, so if both terms appeared in the data set as just the term "woman," then the frequency of the subjects would appear more accurately sized in the word cloud.To complete the translations I used the websites Word Reference and cross-referenced the results to Google Translate.

As I expected, with the titles all translated into English, the words "woman" and "girl" appear a little more predominant. "Old," "garden," and "lady" also appear a little larger. Note that "nos" refers to "Nos." as in "numbers," and is not a foreign language word (e.g. "us" in Spanish).

My data-cleaning is by no means perfect. These visualizations rather serve as an example of all the ways you can manipulate just a single field of textual data to discover patterns. With these simple steps, I can see at a glance that many of the artworks in the 1913 Armory Show featured women, girls, gardens, villages, and heads, because they all appear roughly the same size in this word cloud.