A panel of three papers:
Project Arclight: Critical Reflections on Search, Visualization, and Media History’s Big Data
In this paper, the PIs of “Project Arclight: Analytics for the Study of 20th Century Media”—which received a $200,000 Digging into Data grant sponsored by IMLS, SSHRC, and NEH Office of Digital Humanities—argue for the need to think critically about the data and visualizations used in Digital Humanities projects. Speaking to the conference’s theme of “Art and Science in the Digital Humanities,” we share some of the challenges involving technical engineering and artistic design that we have faced in attempting to develop a software app that is transparent, self-reflexive, AND still accessible to a wide range of users.
Our talk is organized in two parts. First, US PI and former HASTAC Scholar Eric Hoyt briefly describes Project Arclight and talks about the project’s connections to science and technology—especially, search technologies that the Digital Humanities have tended to dismiss. Second, Canadian PI Charles Acland talks about Project Arclight’s approaches to data visualization and engagements, focusing on the project’s collaborations with artists and graphic designers.
The primary goal of Project Arclight is to develop a web-based tool for the study of 20th century American media through comparisons of how entities trend across time and space. As Hoyt explains, the Arclight app analyzes roughly two million pages of public domain publications derived from two repositories: the Library of Congress Chronicling America National Newspaper Program and the Media History Digital Library, a digitization project that Hoyt co-directs.
Hoyt argues that the search technologies have been unfairly dismissed by Digital Humanities scholars, who have called on the field to go “beyond search” and engage with more advanced forms of data mining, such as topic modeling. To adapt search to the scale of big data, Hoyt proposes Scaled Entity Search (SES), which serves as Arclight’s algorithmic backbone. SES allows users to search hundreds or thousands of entities across a corpus simultaneously, and, in so doing, restore the context lost in typical keyword searching. SES balances critical reflection on the entities, corpus, and digital with an appreciation of how all of these factors interact to shape both our results and our future questions.
In the second part of the talk, Charles Acland argues that Project Arclight and other DH projects need to innovate new approaches to data visualization in order to represent adequately underlying rhetorical powers. As Johanna Drucker has pointed out, trending charts and line graphs reify data, forcing them to fit uncomfortably into standardized metrics, and seductively suggest certainty and self-evidential results. Acland will showcase some of Arclight’s new visualization methods, including the Kernel Density Estimation scatter plot, and discuss how Arclight proposes to build more critically attuned visualization structures.
Conference attendees will walk away from this presentation acquainted with Project Arclight and, even more importantly, thinking about how more critical engagements with data, technology, visualization, and design can enhance their own Digital Humanities work.
From Theory to Action: Good Enough Digital Preservation for Under-Resourced Cultural Heritage Institutions
Libraries, archives, museums, and other cultural heritage organizations collect, create, and steward a rapidly increasing volume of digital content. Both research conclusions and professionals’ real-life experiences expose the inherent fragility of this content. The cultural heritage and information science communities have developed guidelines, best practices, policies, procedures, and processes that can enable an organization to achieve high levels of digital preservation. However, these protocols are often complex, leaving many practitioners attempting to address the challenge of preserving digital materials feeling overwhelmed. This is particularly true for professionals serving smaller institutions that are often operating with restricted resources like small staff sizes, a lack of specialized expertise, dated technical infrastructures, and/or limited budgets. This white paper is the result of a three year investigation, funded by the Institute of Museum and Library Services, of affordable, scalable digital preservation solutions that can be successfully implemented at under-resourced organizations. It reports the results of large-scale testing of several digital preservation tools and services, suggests pragmatic digital preservation options, including an incremental approach to digital preservation practices, and asserts that communities of practice are key to success.
The Digging Condition of Digital Humanities: Historicizing the Material Turn through Sound
During the 2000s, a material turn occurred in digital humanities research, with an emphasis on how new media are not ephemeral info-dust. They are inscribed onto platters, embedded in infrastructure, transmitted through wires, and grounded in platforms. Put this way, the material turn responds, if only tacitly, to critical theories of technology, virtuality, and cyberspace that, during the 1980s and 1990s, largely ignored the particulars of how new media actually work. With this turn, we observe what I call a "digging condition" across digital humanities: scholars are now approaching media archaeologically, scraping data, emulating obsolete programs, reanimating dead tech, and unpacking the hidden lives of objects. But, from a historical perspective, how did new media become material in the first place? What motivated, say, the transduction of the ephemeral into the permanent? Engaging these questions, this talk argues that early magnetic recording practices (1878 - 1920) should inform how we understand the emergence of the digging condition in digital humanities. In so doing, it uses sound studies as its foundation, building upon and often complicating tendencies in digital humanities to privilege book history, print culture, and electronic text.