Image 1 and 2: Image Montage created using Montage via ImageJ. Top montage uses the costume design sketches, and the bottom uses photographs from the same collection.
In her final blog post, Doris Morgan Rueda considers the issues and problems that occurred in this initial foray into Collections as Data work with Thomas Padilla, and discusses what she wishes she would have known prior to beginning her research. This exploratory research used over a thousand scanned images of costume designs and photographs that were digitized as part of our LSTA entertainment grant project.
As the Collections as Data project comes to an end, it is helpful to review the frustrating dilemmas that we encountered. As previously mentioned, file naming conventions proved to be a bigger issue that I would have imagined. The process of renaming files to reflect their date pushed us to ask why having chronology was so important. Was it possible to learn something new from the data if we ignored dates for the purpose of visualization? In a larger sense, we were debating the issue of how much focus should be given to context versus an individual item. Should items be looked at individually rather than as a small piece of a larger historical narrative that the collection reveals? Ideally, the process should be somewhere in the middle of these two perspectives. However, there is great benefit to enriching metadata and our understanding of the collection by visiting these extremes.
Image 3: Side by Side comparison of image with their Surface Plot map which show individual data points as opposed to a flat surface.
As a historian, it was difficult to let go of chronology. As we experimented further with different forms of data visualization it became apparent that separating the items from their context often revealed new information about each item and its original context. For example, creating a visual montage of the photographs from the digital Entertainment Collection (which came from a variety of different archival collections), revealed information about where and when the majority of the photographs were taken, which subjects appeared with more frequency, and even which colors seemed to epitomize Vegas entertainment (yellow and red). Looking at the photographs in this way, as a massive whole without any respect to time, origin, or context, gave us the opportunity to form new questions and gather more information that could be used to enrich their individual metadata.
Another problematic issue was the technology itself. ImageJ, developed by the National Institutes of Health, was designed with biomedical researchers in mind, but was able to provide us with the flexibility to do the various data visualization projects we wanted. Being new to the software, it took time to learn, and I often ran into issues when the terminology and intent of the software did not line up with the intent of digital history. Although frustrating, it demonstrates that digital history can and should look beyond its field for technology and ideas about what is possible.
Image 4: Another Side by Side using a black and white photograph and its corresponding Surface Plot map.
While I came to this project as a historian first, this process pushed me to think as an archivist or librarian. Digitization, which has made remote research possible, forces archivists and librarians to make a decision between volume or value. Access is usually thought of as being as simple as making everything digital. However, what good is access to a digital archive if it is not searchable or understandable? Mass digitization without the cataloguing and creation of metadata is therefore no different that traveling to a physical archive. We need to rethink what digital collections are, a collection of items and independently created metadata, as opposed to just a digital archive. Rather than just approach collections with a view as to what they can do for our project, we need to consider what the collection can tell us about itself as a continually evolving dataset.
Doris Morgan Rueda is a doctoral student in the UNLV History Department. Her research focuses on the development of juvenile justice systems in the American Southwest with a special interest in international juvenile justice and race in the twentieth century. Her work uses interdisciplinary methods and a transnational approach to explore the representation and racialization of juvenile delinquency in border towns through legal systems and popular images.