To commemorate the birth of the great American writer Edgar Allan Poe, today's post looks at the content words Poe used most frequently in his writings and then at which songs in Billboard's year-end Top 10 collection (1951-2016) most heavily draw from this bag of words.

The first step in conducting this analysis was to build a corpus of Poe's writings. Like the post on Martin Luther King that I did earlier this week, I used an opportunity sample to build the corpus — in this case, by downloading two volumes from a five-volume set of Poe's writing offered by the Open Culture website. A more thorough analysis would consist of creating a larger corpus and would sample for such variables as year of publication and genre, but here I merely conducted a quick and dirty investigation of a small sample of Poe's language.

In his personal life, Poe seemed to straddle the lines between Old World and New World, North and South (U.S), natural and supernatural, and classical and modern (or, at that time, “Romantic”). And it could be argued that his transcendence of such categories is reflected by the function words he uses, including such shibboleths as thy and shall. At any rate, it is worth noting that after using AntConc to create a word list from this corpus, there is a vast array of function words topping the list, with the first content word appearing at 71st in the list of all words. In order to focus on Poe's highest-ranking content words, several stoplists were applied to the word list, with the resulting list of content words (and an accompanying word search) that can be found by clicking here.

I then ran the word list drawn from Poe's work on my Billboard Top 10 collection by using BRDi Tools (thanks Clayton Darwin and Betsy Barry for assistance in this part of the investigation). The analysis provided scores for chunks of lyrics from the collection and pointed to two songs in the Billboard collection as being most Poe-like, using the bag of words approach. The first is a song by Ed Sheeran (2015):

The high score for this song reflects the romanticism that is front and center in much of Poe's work, as well as references to body parts, which is pervasive in this song as well as in Poe's work.

The second song that the analysis pointed to is this gem from J. Frank Wilson and the Cavaliers (1964), a song that almost perfectly captures Poe's obsessions with both romantic love and the macabre:

This post is intended to show how the tools of corpus linguistics can be used to find similarities between works done in different media. I hope to do more work like this and then present the results in future posts.