An overview of computational analysis of text, foundations, and exploration of challenges and strategies.

Prep

• bring text to play with in Voyant

Going down a rabbit hole: anatomy of a digital book

How is a digital book made?  How does the structure relate to it's functionality? What opportunities does this afford us in terms of text analysis?

(1) Let's consider this book: Alice's Adventures in Wonderland.  Explore the controls on the right side of the page turner.  

• Note the different views of the book. What types of digital files do you think it likely that a digital book is composed of?
• How are the files likely to be made?
• What functions might each type facilitate?

(2) Now consider the controls at he top of the page turner, and explore those. 

• There is a box labeled "search in this text".  What can you deduce about the book from this functionality? 
• The arrows move the view back and forth in the book. The sections allow for jumping to chapter starts.  What must enable this function? (Food for thought...)

(3) What might this page be?  (It also has this view.) Is this also part of the book? 

Diving deeper into text: OCR processes are not perfect.

Consider some areas of special challenge:

• poor image quality (resolution, warp, skew, crop)
• language specific challenges - Arapaho gospel of St. Luke
• books where characters vary - handwritten manuscripts, block printed books
• font styles
• Unexpected arrangement
• Roman/non Roman
• Special character sets: Japanese done well, and not so well
• Mixed languages: Latin/Greek, German/Greek

What are the opportunity points that the structure and arrangement of a book afford?

• Which of the views in HathiTrust are easiest read by humans?
• Which can be read by machines?

Computational analysis of text

We count words - tokenization; why tokenize?

Let's look again at the Arapaho gospel of St. Luke.  Switch to text view.

• What is a word?
• What isn't a word? 
  
  • Punctuation
  • page numbers
  • grammatical constructs of various languages (ex: negation, especially in French; passive reflexive in Spanish)
• What words do we prefer not to count?
  
  • headers/footers
  • preface
  • indexes
  • title pages
  • abstracts

Introducing Control - "Microanalysis" and Voyant

• Voyant
• Load your text sample
• Playing with control
  • observe counts
  • stopwords
• What does exerting control do to our results?  Does it change the validity of our assertions?
• What is signal?  What is noise?

We calculate frequency

• Why not express our counts simply?  Why calculate frequencies?
• Is this misleading?  If so in what ways?

Moving from Microanalysis to Macroanalysis

• Google nGrams; help (examples)
• Bookworm; help, (examples)
• problems of convergence/divergence and strategies for disentangling.

More macroanalysis

• HTRC
• entity extraction
  • What are entities?
  • False positives/false negatives (omissions)
• Clustering - topic modeling
  • define and defer for Mimno

Image analysis

• Edge analysis
• Neural networks

Biblio

Writing about your results - the structure of scientific papers

Matt Jockers Book