Before the institute and today’s session, I thought that I might be able to use text mining to aid me in my research. More specifically, I had intended to use it to examine a cluster of apx. 75 oral histories on riots that took place in Baltimore following Martin Luther King, Jr.’s assassination. These interviews have already been transcribed and are readily accessible at the Baltimiore 68 website. However, after doing the reading and playing with some of the search engines, I am less sure about the usefulness of this technique for my project. To begin, it appears that “distant reading” is best for large bodies of work, and while 75 transcripts, of apx. 15-20 pages each is significant, I’m not sure if the text mining enhances my ability to interpret the “data.” In addition, I’m not sure what the results of some of the mining shows. For example, the word clouds and/or charts generated by voyant either don’t say much or reveal things about the sources that I didn’t already know. This said, I intend to continue to experiment with overview and voyant to see if they reveal some patterns that otherwise are not readily apparent or which I have not seen.
Source: Distant Reading and My Research