In Unit 3 we’ve been working in R to measure features that are more complex than

In Unit 3 we’ve been working in R to measure features that are more complex than word counts, looking at document-term frequencies and relationships among words. Accordingly, Project 3 will use these methods in R to ask a question about a text or a series of texts, finding new insights to which we otherwise wouldn’t have access.
For this third project, please choose a medium-to-long text (e.g., more than 2,000 words) or collection of texts (e.g., a collection of stories, poems, a collection of chapter, as long as it’s longer than 2,000 words) that you would like to study and use R to answer some non-obvious question about it. Unlike with previous projects, this report will be delivered as a presentation, so prepare slides with five of the six sections outlined in the guide to writing project reports (minus an abstract). The resulting presentation should be between 6 and 10 minutes long, and it should have at least ten slides with speaker notes. The report will be graded based on the rubric.
The question you choose to study might vary greatly in scope and specificity. In the same way you’ve chosen your own text or set of texts, you should also choose your own question to answer something that drives your interest. The options are broad, with two caveats:
Ask a question you haven’t asked before, about a text / set of texts you haven’t studied before.
Ask a question about a text that wasn’t directly worked through during one of the explanations offered by the instructor.
Here are some sample questions to give you some ideas, with parenthetical additions suggesting the kind of thing that would bear further consideration in the Discussion section of the report:
Romance novels share a lot of things in common, but they also show important differences. Using tf-idf, what are the words that distinguish mystery novels from each other, and how do these distinctions give insight into the topics being considered? (And in what ways can this information be more limiting than simply studying word frequencies?) Note: You might need to filter out character names for this approach to be insightful.
The Federalist Papers was an attempt to persuade 18th-century New Yorkers of the merits to be found in the proposed Constitution of the United States of America. Each of the 85 papers is written anonymously by Alexander Hamilton, James Madison, or John Jay, who signed the letters with the shared pseudonym “Publius.” Using tf-idf, what can you identify as the distinguishing topics for all or some of these documents? (And what, if anything, can this approach reveal about the topic choices of one author?) Note: Since there’s so many, you might find it helpful to limit your consideration to a subset of papers, providing you’re still getting a large enough set of texts.
Consider the major individuals named in one of the two special counsel reports linked on Canvas. Which words are the most common words following, for instance, “the President,” Jones, Lewinsky, Currie, Jordan, or other individuals of interest in the Ken Starr report on President Clinton? Or which words are the most common words following “the President,” Cohen, Manafort, “Trump Jr.”, Flynn, or other individuals of interest in the Mueller report on President Trump? (And how does comparing these connections offer insight into the kinds of things most often said about these individuals?) Note: for some of these individuals, you’ll have to look at multiple words in a row, like “the President.”
Considering the writings on The Church on the Southern Black community, which words most commonly precede or follow a set of words you’re interested in? For instance, which words come before or after words related to an individual’s place in society (country, policy, government, society, citizens, etc.), or which words come before or after words related to significant topics of Christianity (God, grace, sin, forgive, etc.)? (And how do these connections allow us to understand the ways these writings differ among the different ideas shown by your analysis?)
What are the most common bigrams or trigrams of a text or set of texts that you’re interested in? (And what kinds of “distant reading” insights do they offer?)