Using the corpus you brought—or, if you could not attend, a three-to-ten-text set drawn from Project Gutenberg, HathiTrust public-domain holdings, or your own files:

• Create a fresh Project in Claude.
• Upload all texts to the Project’s knowledge.

Keep the set small enough to reason about, but large enough to compare across.

Run the analytical sequence one step at a time, keeping each result in the conversation so later steps can build on it:

1. Stopword filter — strip function words so content words surface.
2. Bag-of-words — raw term counts across the corpus.
3. Key phrases — multi-word expressions, not just single tokens.
4. Character or theme network — what (or who) connects to what (or whom).
5. Comparative read — how the texts diverge and converge across the set.

Ask Claude to generate Artifacts that visualize the findings. We will produce at least five different artifacts and iterate on the prompts and choices in between:

• Word clouds
• Frequency charts
• Network diagrams
• Comparative tables
• Thematic timelines

Iterate. Change the stopword list, regroup categories, re-scope the comparison, and watch how each visualization shifts in response.

• Critique what you see. What does the visualization foreground, and what does it bury?
• Compare to close reading. What would you have noticed reading these texts yourself that the distant view misses?
• Document at least one place Claude got it wrong — a miscounted term, a hallucinated relationship, a flattened theme, or a misattributed quote.

Suggested reading:

• Underwood, Ted. “A Genealogy of Distant Reading.” Digital Humanities Quarterly 11.2 (2017).
• Walsh, Melanie, and Maria Antoniak. “The Goodreads ‘Classics’: A Computational Study of Readers, Amazon, and Crowdsourced Amateur Criticism.” Post45 7 (2021).

Sample assignment:

• Salter, Anastasia. “Week Four: Reading Across Texts.” Humanities & AI.