By Kat Stefko
In a January 4, 2016, special report on Digital Humanities in Libraries, published in American Libraries Magazine, Stewart Varner and Patricia Hswe posited that “Stripping digital collections down to core components could render everything old new again in terms of what libraries might offer to the humanities research community.” Taking this approach to heart, the George J. Mitchell Department of Special Collections & Archives at Bowdoin College Library has been exploring new ways of leveraging metadata about our digital and physical collections to support a burgeoning interest in the digital humanities and computational studies across the College’s curriculum.
Historically at Bowdoin, perhaps like all libraries, we have regarded metadata first and foremost as a functional tool — the description necessary to assist researchers in their inquiries and librarians in their management of collections. Digital humanities, data curation, and new technologies, such as open source data visualization software, prompt us, though, to consider our metadata in a new light. Metadata, considered more abstractly, is one of our most valuable and important collections. Our metadata reflects decades of work by dedicated staff and volunteers, who applied their energy and expertise to analyze, synthesize, and interpret our physical and, more recently, our digital materials.
Over the past year, Special Collections & Archives staff have been integrating metadata in our teaching and outreach efforts. Along with promoting our rich collections of rare books, manuscript and archival holdings, we are actively looking for opportunities to suggest and support the integration of our metadata in faculty teaching and research.
These efforts have coincided with and been expedited by the largest digitization project yet undertaken by Bowdoin—the Howard Digitization Project. Funded by the National Historical Publications and Records Commission, this project allowed us to digitize the entire archive of General Oliver Otis Howard and his two brothers, Charles and Rowland, all of whom graduated from Bowdoin and served in the Civil War. The project produced some 180,000 digital images representing more than 80,000 physical items. Another key aspect of the project was the modernization of metadata about the collection.
Volunteers had thoroughly indexed the incoming letters to Oliver Otis Howard over the course of several decades. Project staff were able to scan, OCR, normalize, and then augment this information to create a comprehensive dataset about the Howard letters. Helpful for managing the digital files, this dataset also offers rich opportunities for the application of digital humanities methodologies and tools. The index includes more than 40,000 records with data points such as the letter’s author, recipient, date, and the place where it was written and received.
In their article, Varner and Hswe asserted that the “representation of digital collections in various data formats may lead to creative programs and partnerships for instruction, collection development and strategy….” And, that has indeed been the case at Bowdoin.
The Howard correspondence index has quickly found a place within the College’s curriculum. In Fall 2016, Clare Bates Congdon, Visiting Associate Professor of Computer Science, had students in her Interactive Data Visualization course work with this rich data set. The resulting visualizations, while simple, are compelling. Researchers, for instance, can select a date range to then see a map of Oliver Otis Howard’s correspondence for that time period. By expanding the chronological range, a researcher can quickly see the arc of Oliver Otis Howard’s career as reflected by his correspondents. The visualization shows a clear southward and westward expansion of Howard’s social network as he transitioned from Civil War general to commissioner of the Freedmen’s Bureau and a major figure in Reconstruction to a participant in the Indian Wars.
While learning new technologies and the ingredients for a successful interactive data visualization, the students also were exposed to the biography of a compelling 19th century Bowdoin graduate. According to Congdon, working with Bowdoin-specific data made the exercise more meaningful for the students and encouraged them to engage with the assignment on multiple levels. The exercise also established a collaborative and mutually beneficial relationship that will continue between the Computer Science Department and Special Collections.
The Howard correspondence visualization project, along with other examples of faculty utilizing metadata from Special Collections, was the topic of a recent faculty development lunch and learn. Organized to mark the conclusion of the Howard digitization project, the well-attended event proved highly successful, and several faculty members have subsequently indicated an interest in incorporating metadata from Special Collections into their teaching and research. Current and future projects include visualizing the personal library of one of the College’s founders based on our bibliographic records of his collection, mapping the travels of a nineteenth-century female botanist based on our index of her plant samples, and experimenting with textual analysis to explore transcripts of nineteenth-century Bowdoin student letters to see what this might reveal about what a liberal arts education offered 150 years ago.
Kat Stefko is Director of Special Collections & Archives at Bowdoin College and current chair of the College & University Archives Section of SAA. She has held library and archives positions at Duke and Harvard universities, Bates College, and the Philadelphia Museum of Art.