Textmining allows you to use computer software to conduct an extensive quanitative analysis of text. Textmining is most useful in conducting an analysis of large corpora in a database or archive. Specifically, textmining can be used to search through text to quanitify how often a certain word appears in a digitized archive or material. Textmining tools can be used to examine patterns or trends in qualitiative data. Several projects that have utilized textmining include America’s Public Bible, Robots Reading Vogue and Signs40. In the case of America’s Public Bible, textmining was used to perform an analysis of biblical quotations in US newspapers to determine trends in biblical references at different historical periods. Robots Reading Vogue was designed to evaluate beauty trends over time while Signs40 used textmining to determine changes in Signs’ feminist scholarship. Trends related to certain words, topics and themes can then enable a researcher to draw certain conclusions on the historical time period of study.
Free text mining tools like Voyant can be used to conduct textual analysis. For instance, you can upload links of textual information to Voyant as a dataset that will then be processed into a digital corpus. From there, tools like Cirrus, Reader, Trends, Summary and Contexts can be used as visual interactives to find patterns or differences in the data. The dataset used for my exploration of Voyant was the WPA Slave Narratives.
I utilized all of Voyant’s given tools to compare the frequency of words in various states. Specifically, I compared the states of Alabama and Maryland to each other by seeing the distinctive words utilized in their interviews. For instance, Alabama used distinct words like didn’t (357), alabama (138), ain’t (213), don’t 182), ca’se (69), caze(104), couldn’t (103), i’s (93), i’se (92), dat’s (90). Maryland used distinct words like bowie site. (17), baltimore (64), ellicott(11), cockeysville (11), arundel (9), annapolis (9), md (11), maryland (60), rezin(7), manor (6). From this analysis, I was able to conclude that Alabama’s distinct words referred to their unique dialect while Maryland’s distinct words referred to Maryland African American communities. In using Voyant, I was able to see that while textmining is useful in searching through large sets of data-it is imperative that a person is working in conjuction with the software to get rid of unnesessary “stop” words and to make sense of the words that are repeated throughout to draw historical conclusions.