A space to share ideas and experiment in digital humanities

Humanities scholars are accustomed to closely studying novels, essays, archival records, and other sources. They question the material. They draw conclusions from it and wrestle with big ideas. Increasingly, they also wrestle with big data.

Google has made more than 30 million books available online. Digitized runs of newspapers and magazines are available through subscription services. Art museums, such as the Yale Center for British Art and the Yale University Art Gallery, are digitizing their collections, as are major libraries, including Yale’s Beinecke Rare Book & Manuscript Library and the Library of Congress.

This explosion of digital material presents both opportunities and unique challenges to humanities scholars, who traditionally have not used computational methods to manage massive amounts of data.

Yale humanities scholars — faculty, graduates students, and undergraduates — have a place to turn for help when confronted with gigabytes of data: the Digital Humanities Lab at Sterling Memorial Library.

The lab, which opened in the fall of 2015 with the support of the Goizueta Foundation, facilitates collaboration between humanists and scholars of science, technology, engineering, and mathematics (STEM). It provides scholars support and resources for projects that apply techniques mastered by Silicon Valley stalwarts like Facebook and Twitter to troves of humanities-related data, said Peter Leonard, the lab’s director. 

“We’re not applying big data techniques to people’s tweets, but to 19th-century novels; not to photos uploaded on Facebook, but to a digitized archive of photographs from the Great Depression,” he said.

The work does not simply involve humanities scholars seeking help from their colleagues in STEM fields, Leonard said. It is a matter of sharing expertise and collaboration.

“Humanities scholars realize that only big data techniques can make sense of this massive corpus, but they know that they can’t just rely on a computer scientist to solve this problem for them in every case,” he said. “The smartest computer scientist, the best applied statistician, the top engineer at Facebook — he or she is not going to know the questions that are relevant to an English professor at Yale.  

The English professor is going to have unique questions that are grounded in his or her own theoretical perspective.”

The lab has a web designer, programmer, and outreach coordinator on staff to provide advice, workshops, and consultations at varying levels of formality.  It also has equipment that scholars can use to digitize material from Yale’s world-class collections to create their own digital archive to work with. It is a place to share ideas and experiment, Leonard said.

“What we’re doing is building a lab that fulfills some of the roles labs play in the hard sciences,” Leonard said. “We’re not trying to argue that the laboratory model is always appropriate for the arts and humanities. We’re not trying to change the way all of this work is done, but we are saying that there are some benefits to having a central resource for doing intensive data or algorithmic work.”

The lab also distributes grant money. It provides grants of $3,000 or less to graduate students and faculty for start-up projects. Each year, two $20,000 grants are awarded to Yale faculty for humanities-related digital projects via a competitive process overseen by the 14-member Digital Humanities Committee, which is composed of faculty and staff members and focuses on advancing digital humanities broadly at Yale.

The latest $20,000 project grants were announced on Nov. 30. Matthew Jacobson, professor of African American studies, history, and American studies, received a grant to support two projects — a documentary and a digital game — concerning life in the South during the Jim Crow era.  The other grant recipient, assistant professor of music Anna Zayaruznaya, is creating an interactive platform to study and engage with the Roman de Fauvel manuscript, a famous 14th-century satirical poem.

Grant recipients spend the money according to a project plan that they develop using guidelines modeled after the grant application for the National Endowment for the Humanities.

“They might use it to hire graduate students to supervise a process of digitization or data scrubbing — there’s a lot of unglamorous work to do — or they might want to hire a programmer,” Leonard said.  “We trust they know their problem. We’re here to provide advice, to connect them with a programmer and to consult with them on how to best structure a big digital project.”

Photogrammar

The U.S. Farm Security Administration and Office of War Information (FSA-OW) photo archive consists of roughly 170,000 photographs documenting American life from 1935 to 1944.

The product of a photographic project by the Roosevelt administration to build support for its federal aid programs, the archive features some of the most iconic photographs of Americans taken during the Great Depression and World War II. It includes the work of several masters of 20th-century photography, such as Dorothea Lange, Walker Evans, Ben Shahn, and Arthur Rothstein.

The archive is housed at the Library of Congress, which digitized it and made the images available online, though the massive collection is difficult to explore effectively through the library’s search tool.

Photogrammar, the Digital Humanities Lab’s first major project, has made it much easier for scholars and the public to navigate and make sense of this enormous, unwieldy, and extraordinary archive. It is the product of cross-disciplinary teamwork, as well as skill and creativity.

Laura Wexler, professor of American studies and a scholar of photography, is the principal investigator for the Photogrammar project. She knows firsthand the difficulties of exploring the archive. She was among a group of historians invited by the library in 2001 to search the collection for hidden masterpieces.

“Even a group of six or eight scholars going through the physical archive for several days couldn’t search efficiently,” said Wexler, who has a joint appointment in American studies, film and media studies, and women’s, gender and sexuality studies.

About six years ago, Lauren Tilton, a graduate student in American studies, was working on an assignment for Wexler on the FSA-OWI archive when she began thinking about creating a better way to study the collection.

Tilton connected with Taylor Arnold, a graduate student in statistics, to help her create a tool for searching the archive. They developed a National Endowment for the Humanities (NEH) start-up grant proposal and turned to Wexler for faculty guidance.

“It grew from there,” Wexler said.

In 2011, they were awarded an NEH grant — the first NEH grant to fund a large-scale digital project at Yale — to create an interactive map that would allow researchers to better organize, search, and visualize the FSA-OWI archive.

Wexler and co-directors Tilton, and Arnold assembled their team: Peter Leonard; Trip Kirkpatrick, senior academic technologist at the Yale Instructional Technology Group; Stacey Maples, former GIS specialist and instruction coordinator at the Yale Library; and Ken Panko, at the time the director of Yale University's Center for Media and Instructional Innovation.

“The strength of our team was very important to me,” Wexler said. “I recognized immediately that this was a brilliant group of people with an incredible project. I felt we could be an example of how this sort of interdisciplinary team could work at Yale — combining computer science, statistics, library science, and humanities.”

Photogrammar’s interactive map plots about 90,000 photographs from the FSA-OWI archive onto a map of the United States.  Users can follow the paths of individual photographers to discover where they visited. They can overlay a 1937 Vico Motor Oil road map of the United States to chart the journeys of the 12 most notable photographers. They can delve deeply into the associated metadata.

Wexler said the interactive map “destroyed the truism” that the FSA-OWI photography project focused entirely on the plight of Americans in Appalachia and the Dust Bowl.

“It turns out that is wildly inaccurate,” she said. “The moment we mapped it, we learned that the photographers’ work extended throughout the country and into the cities.”

Photogrammar provides other tools for visualizing the archive, including a three-tiered tree map that categorizes 88,000 photographs by 12 subject headings (e.g. “Work” or “The Land”), 1,300 sub-headings (e.g. “Farms” or “Mountains”), and sub-sub-headings (e.g. “Barns” or “Drought Conditions”.)  The tree map is based on a classification system designed for the archive in 1942 by Paul Vanderbilt, a librarian, curator, and photographer.

The Photogrammar team recently secured another grant to fund its next phase, this time from the American Council of Learned Societies in their inaugural round of digital extension grants. New projects will include the addition of oral histories of the photographers and a new feature that will digitally reassemble the photographers’ rolls of 35mm film, which were cut into strips of about five to six frames apiece for processing, Wexler said.

(Tilton and Arnold have completed their PhD. programs and have left Yale to take faculty positions at the University of Richmond, but they remain co-directors of the Photogrammar Project.)

Wexler, a trained photographer, said that the ability to view the exact sequence of shots creates a window into a photographer’s thinking. For instance, the Photogrammar blog foregrounds a sequence of photographs in which John Vachon took several shots illustrating the cheese industry in Wisconsin and then moved directly to shots of a homeless man in Chicago.

“Why the jump to a homeless man in Chicago?” Wexler said. “His question here is this: We have cheese in Wisconsin. Why can’t we get it to Chicago where hungry people need food?”

Wexler, a member of the Digital Humanities Executive Committee, said the Digital Humanities Lab is vital to Photogrammar and other major digital humanities projects.

“If we didn’t have a Digital Humanities Lab; if we didn’t have people saying, ‘Oh that’s interesting, let’s see what you can do’; and if we didn’t have passionate people willing to do the work, then Photogrammar wouldn’t have been built,” she said. “Peter has made the Digital Humanities Lab such a friendly space for all of these things to come together.”

Wexler likes to hold her office hours at the lab, which is currently located on the third floor of Sterling Memorial Library. (It will eventually move into the Franke Family Reading Room near the library’s main entrance.)

“I like to be there because so many things are being discovered and shared in that space,” she said. 

Conviviality and intellectual curiosity

Elihu Rubin, associate professor of architecture and American studies, has spent a decade working with his students to amass an archive of information about the architecture and streetscapes of New Haven, as well as the people who have lived and worked in the city’s buildings.

With support from the Digital Humanities Lab, including a $20,000 project grant, Rubin is building an interactive web platform like Photogrammar that allows people to efficiently search this archive of New Haven buildings, visualize how the cityscape has changed aesthetically and demographically over time, and spark conversation about urban planning and development as well as memory and place attachment.

Rubin, also a member of the Digital Humanities Executive Committee, said Leonard and the lab’s staff have created a warm and open atmosphere that invites creative thinking.

“The environment allows for intellectual connections,” he said.  “It’s a place where computer science, interaction design, and humanistic storytelling can come together.”

Marijeta Bozovic, assistant professor of Slavic languages and literatures, affiliated with film and media studies, and women’s, gender, and sexuality studies, received $20,000 project grant from the lab to support Digital Humanities and Russian and East European Studies (DHREES), a web-based platform she is creating to explore intersections between digital humanities and Russian and East European studies.

The multi-faceted project concerns digital approaches to literary and cultural studies, interrogating continuities with Russian Formalism and other critical theories, as well as finding intellectually stimulating ways to use technology to enhance teaching and learning. For example, Bozovic is working in partnership with colleagues at Columbia University on the Black Sea Network — an initiative to share resources across institutions and nations.

One of the search tools the DHREES team has built is a topic browser of all the issues of the academic publication Slavic Review available through JSTOR, an online repository of scholarly journals. The search tool identifies recurring and dominant themes in the journal’s subject matter, such as cinema and the media, gender and sexuality studies, or post socialism.

“We can see when some of these topics entered the academic discussion,” Bozovic said. “When are they dominant? When are there dramatic changes in the academic discourse about Slavic studies or Russian and East European studies?”

The lab has focused on the vast amounts of digital material the Yale Library licenses, including academic journals, magazines, and newspapers, Leonard said.

“This material oftentimes is very meaningful to humanists,” he said. “It could include a run of old newspapers from the Caribbean or The Times from London. It could contain archives of a counterculture newspaper from the 1960s or photographs from the Civil War — all sorts of things that vendors license to Yale. We work to make this digital material amenable to the same types of research one might perform by taking a book from the library.”

Leonard and Lindsay King, associate director for access and research services at the Robert B. Haas Family Arts Library, collaborated on Robots Reading Vogue, a project that facilitates all sorts of data-based experiments on the complete run of Voguemagazine, encompassing 2,700 covers and 400,000 pages.

Rubin noted that digital projects like Photogrammar and Robots Reading Vogue require long-term stewardship.

“That’s a big challenge,” he said. “Websites are always going defunct. Increasingly, valuable scholarship will appear in digital formats, not conventional ones. The Digital Humanities Lab has a critical role to play in curating these projects over the long term.”

Wexler said Yale’s efforts to attract more students interested in STEM-related fields will naturally drive digital humanities collaborations, heightening the importance of the lab as space where people can meet, brainstorm, and launch new projects.  

“Frankly, I think we have fallen through the lens of the camera into the digital realm,” she said. “We’re not coming back from the digital world.”