The Office of Digital Humanities at the National Endowment for the Humanities (NEH) has awarded a grant to fund Yale’s Photogrammar Project, an interactive, interdisciplinary website that will make some 170,000 historic images accessible to a new generation of scholars and educators.
The photographs, taken between 1935 and 1943 to document federal aid programs during the Great Depression and to address the morale of the nation at the beginning of World War II, were a project of the Farm Security Administration (FSA) and the Office of War Information. Many images show the hardscrabble lives of Southern sharecroppers and migrant farmers taken by such legendary photographers as Walker Evans, Dorothea Lange, Ben Shahn, and Gordon Parks. Others depict America’s efforts to prepare for World War II.
Graduate students Lauren Tilton (American Studies) and Taylor Arnold (Statistics) are working to create the Photogrammar Project, an initiative of Yale’’s Public Humanities Program and Photographic Memory Workshop.
The project is directed by Laura Wexler, who has a joint appointment in American Studies and Women’s, Gender & Sexuality Studies and is founder and director of the Photographic Memory Workshop, a cross-disciplinary working group of Yale faculty, staff and students that explores the intersection and interplay of photography and memory.
“The Library of Congress has digitized these invaluable photographs, along with available information about who took them, when and where, and that is a crucial first step for any digital public humanities project,” says Wexler. But, she adds, “Our project creates new ways to search the database through an interactive map that will plot the photographs over historical county, state, national and census data over time. We also plan to associate the photographs with other kinds of material, such as the famous ’shooting scripts’ of the FSA and the papers of its director, Roy Stryker.”
Lauren and Taylor initially wrote a computer program to download the photographs and metadata (such as the photographer’s name and the date and location of the photograph) from the Library of Congress American Memory website. They next created a “proof of concept” website using a small subset of the collection and tools from Google maps. “With only this smaller collection of images and limited graphics, we have already discovered some promising new patterns that we are excited to see on a larger scale,” says Lauren. “Our project will allow researchers to back-up, or even challenge, previous positions about the archive and the period of history it recorded with direct visual and quantitative evidence, while discovering new patterns that would otherwise be undetectable by simply going through the photographs one by one.”
With funding in hand, they are now starting to obtain historically accurate maps and relevant census data from the Map Collection at Sterling Memorial Library, aided by information specialist Stacey Maples, and to build a robust open source web platform, working with Ken Panko from Yale Instructional Technologies.
“We need to create our own mapping module to accommodate the size of our dataset, since Google Maps does not work for such a large set of images,” says Taylor.
“Users will be able to construct statistical graphics and visualizations from the data. For example, a user will be able to quickly plot the percentage of military images collected by month and location or see a gallery of sharecropping images created in Georgia,” Lauren explains.
Lauren became involved in the project as an outgrowth of a course assignment for Professor Wexler’s “Introduction to Public Humanities” seminar, which she was taking for the master’s concentration in public humanities. Her academic interests include studying the portrayal of race and poverty, specifically the changing narratives that the photographic archive reflects as the nation moved through the New Deal era and prepared for World War II. Lauren came to Yale with the intent to study visual culture and photography under Wexler. She expects that her dissertation will focus on documentary expression and visual culture in the 1930s and 1940s. A serious amateur photographer, Lauren has had several pictures published in Southern Spaces, an online peer-reviewed journal dedicated to the study of the American South.
Taylor’s involvement arose from his interest in textual analysis of the digital humanities (i.e., using computers to study a large corpus of texts) and the unique opportunity the project gives to apply statistics to the visual studies aspect of the digital humanities.
“A large part of the emerging field of digital humanities is concerned with analyzing and visualizing large datasets like our collection of images, and yet has not garnered much interest from statisticians,” he says. “I am hoping to rectify this, while at the same time deal with a collection of data that few, if any, other people in my field have been able to explore.” His dissertation, advised by John Emerson and Sekhar Tatikonda, deals with detecting low-dimensional structure in high-dimensional datasets. This involves, for example, taking a dataset of 20,000 genes and determining which five are actually involved in a particular biological process. “My particular interest is turning theory into algorithms that can be applied to study real datasets,” he says.