I am currently collaborating as co-PI with Amber Stubbs (Computer Science) on the development of a database of digitized court records digitized from eighteenth and nineteenth century slave jails in North America. The records are drawn from archival collections, published volumes (with permission) and online repositories such as the Internet Archives and HathiTrust.
This project began with research at the Library of Congress under the auspices of the 2018 NEH Slavery and the Constitution Seminar run by Paul Finkelman and Paul Benson. That fall, my research led to a collaboration with Amber Stubbs, in the Simmons University Department of Computer Science.
The prototype of the database, which uses MySQL and PhP language, was built as a collaboration with students, funded by a grant from the Computer Research Association for Women and the Institute for African-American Mentoring in Computing Sciences. In spring 2019, I mentored the students on reading historical documents, research, and working with metadata in a humanities setting while Amber Stubbs supervised the building of the database. The students presented the first working prototype of the database at the Undergraduate Research Symposium in April 2019.
Another small grant from Simmons University designed for faculty-student collaborations allowed me to hire an undergraduate student in fall 2019 to spend a semester working through collections in HaitiTrust and the Internet Archives to identify additional court records. I also worked on refining the metadata and supervised additional undergraduate students (working for course credit) on digitizing and OCRing a set of volumes we received permission to digitize. Using a Python script, Amber Stubbs began scrapping the OCR’d records in order to run some natural language processing so that we could improve our topic modeling. This project has also drawn from work by Daryle Williams, Jessica Lu, and Caitlin Pollock on the creation of linked data for projects involving enslaved people.
The database is still under construction, but the current prototype can be found here. A screenshot of our dataset from the prototype appears below