DocSouth and SEASR

Title of Study/Project:

List of team members and their affiliation

Natasha Smith, UNC-Chapel Hill, DocSouth collections

SEASR Contact

Loretta Auvil

Procedural Outline of Study/Project

Research Question/Purpose of Study

Temporal Coverage - great interest and use for browsing across a large body of content (texts, as well as collections)
Textual analysis
Places and time

I have two PhD candidates who are interested in working with us. A few suggestions/thoughts/ideas from them:

" The tag cloud option might be an interesting application to offer,
either on a search page or as an option from the home menu for each
text."
"The format for the search page or application could/should be styled
after the Google Book Search, which gives a snippet view of the lines
just before and just after the requested word or phrase. For an
example, please see the following link and type "three feet" in the top
right search field:
http://books.google.com/books?id=VNK0QUMnoCkC&printsec=frontcover&dq=harriet+jacobs+incidents#PPA173,M1
"This kind of text analysis raises an interesting question about how
well the frequency of word use correlates to important themes or
concepts in the text. Knowing what I do about Douglass's texts, I am
not surprised to see "man" "slavery" and "slave" appear prominently. I
also think it is interesting that the word "colored" is more prominent
in the 1892 version than the 1845, since it suggests visually that an
important shift in terminology has occurred between 1845 and 1892.
Because these images are just word count results that ignore context,
it's impossible to use them as proof of an idea, but they do provide an
interesting and suggestive jumping-off point that could suggest any
number of research questions that could be answered by looking at the
words in their context.
It would be most useful if this presentation was an initial page which
could then lead to a more detailed break-down of how often each word
appeared and possibly a link to a brief (sentence or less) snippet of
each occurrence so that a scholar could easily see how a word was being
used."
"It would also be nice if there were a variety of levels of filtering
that the user could implement. For instance, one person might be very
interested that the word "Mr" appears so often, while another might
view it as an unnecessary common word, like "the". Likewise, a user
might want to eliminate the word "page" from the visual result because
it is (I'm assuming) related to the page numbers rather than the
content of the text itself. I think a declaration of how the filtering
works would also be necessary -- I found myself wondering what criteria
the program used to filter out words like "and," "the," and "is"."

Data Source

Suggested titles for text analysis:

different editions of a narrative from Slave narratives
1. F. Douglass - 1845 first (125 pages)and 1892 last (752 pages) editions
2. Moses Roper - 1838 edition and/vs 1848 edition

Suggested titles for work on exctracting dates/temporal coverage:

Daniel Alexander Payne, 1811-1893 "History of the African Methodist Episcopal Church. Nashville, Tenn.: Publishing House of the A. M. E. Sunday School Union, 1891.
James T. Haley. Afro-American Encyclopaedia

Specific suggestion from a very known scholar Prof. William Andrews who served as the scholarly advisor on the Slave Narratives project:

He suggested to look into places/time of a world travelor/slave Olaudah Equiano. His life and summary of his 2-volumes accounts The Interesting Narrative of the Life of Olaudah Equiano, or Gustavus Vassa, the African. Written by Himself.

Analysis Tools

For temporal coverage - Loretta mentioned a "dates simile"
Text analysis - changes in editions of the same work?
"summarizer"?
words count
most commonly used
frequent patterns

Activity Timeline or Milestones

NEH SUG application http://neh.gov/grants/guidelines/digitalhumanitiesstartup.html- October 6, 2009:

need to identify WHAT we need to plan to build/work on