Innovative Research and Creative Output: From Ideas to Impact
/Society for Scholarly Publishing - Philadelphia Regional Event
This post is a collaboration between SSP members, including Nicola Hill, Emma Sanders, and Adrian Stanley.
Left to right: Kathi Martin, Drexel Digital Museum; Jen Grayburn, CLIR Postdoc; Alex Humphreys, JSTOR Labs
On October 30th, the Society for Scholarly Publishing (SSP) hosted a regional event at the University of Pennsylvania, Van Pelt Library. The topic, "Innovative Research and Creative Outputs: From Ideas to Impact" brought together Philly-area publishers, librarians, and content professionals for a panel discussion on new and innovative methods of producing scholarship.
Jen Grayburn, CLIR Postdoctoral Fellow
Jen spoke about her use of Google Scholar, SketchFab and Unity in her work, which centers around the intersection of architecture and text. Using GIS (Geographic Information Systems) mapping software, Jen examines locations of historic sites. She shared an example of a mapping she did of St. Magnus Cathedral in the islands off the north coast of Scotland. In this particular example, Jen generated a binary map that indicated what would and wouldn’t be visible on the ground from a certain height.
She uses geo-TIFs (TIF files encoded with geographical coordinates) to create a 3D topographic map to illustrate what is visible and why. Eventually, these mappings were confirmed with on-site visits she conducted. In her work, Jen uses Sketchfab to store the large 3D modeling files
Currently, there is a lack of standards around 3D scholarly outputs—how they’re reviewed, stored, and made accessible.3D collections are siloed by institution—there is really no repository. The only exception Jen cites is Duke University’s MORPHO SOURCE. For these reasons, evaluating and citing digital work is still a challenge.
Studies in Digital Heritage content are inextricably linked to the 3D model created in the course of those studies. There is a real need for community standards for 3D data presentation. Academic departments are generally slow to reward digital projects, or have a process for incorporating these scholarly outputs in formal evaluations.
Archeologists with an interest in Jen’s work, for example, always want the original 3D model she created, not the version on Sketchfab. But these models haven’t been peer-reviewed, and for that reason, Jen is reluctant to provide. In the near future, more standard development and community standards for 3D and VR creation and curation in higher education is certainly warranted.
Kathi Martin, the Drexel Digital Museum Project
Kathi Martin presented her work with The Drexel Digital Museum Project: Historic Costume Collection (digimuse)---a searchable image database comprising select fashion from historic costume collections. Initially, fashion images were highly protected by using low-res images and watermarked images on the website. Kathi explained that Polish hacktivists demonstrated to her how easy it is to remove the watermark and improve resolution.
The museum has always been driven by open access and open source to share information and further usage and research. Interoperability is key to the museum’s mission—this allows the data on the museum’s website to be easily harvested across browsers.
The museum has widened beyond Drexel’s collection; for example, Iris Barre Apfel’s Geoffery Beene collection was displayed and that exhibit is archived on the museum site. Quicktime VR was used to film the collection and provide high-resolution captures of the fashion collections.
The technology DigiMuse is used in the Drexel project and provides a new level of engagement with the collections Kathi is preserving. Drexel's Digital Museum project website allows a site visitor to interact personally and actively with a distributed, collected narrative. The site includes rich metadata descriptions for every picture. The variety of contributions on the site, Kathi feels, stimulate varying and often deeply personal reactions.
She believes the site is very powerful due to its “baked-in connectedness.” Kathi closed with Grace Kelly’s gown, made by Givenchy in part out of actual coral (gasp!). The site complements the high-res images of the gown itself with media of Grace Kelly in the gown.
Alex Humphreys, JSTOR Labs
Alex discussed how JSTOR Labs applies methods and tools from digital scholarship to create tools for researchers, teachers, and students "that are immediately useful – and a little bit magical." JSTOR is a member of ITHAKA, a non-profit devoted to digital sustainability.
Alex Humphreys, director at JSTOR LaBs
Alex works with a team of five on innovative projects that benefit humanities scholars. He demonstrated JSTOR Labs’ Understanding Shakespeare tool, which uses the Folger Shakespeare Library’s digital version of Shakespeare plays to hyperlink each line of the play to a search showing all JSTOR articles that contain a particular line of prose.
JSTOR Labs works from a philosophy of play—Alex sees what resources other organizations (like Folger Shakespeare Library) bring, what LABS brings, and what kind of sandbox they might build in collaboration. Part of JSTOR Labs’ philosophy values what Alex calls “multi-disciplinarity.” For example, JSTOR Labs’ partnership with Eigenfactor (which measures influential and highly cited articles) has resulted in a tool that helps scholars discover the most influential articles in a given field or topic area.
JSTOR Labs also believes in hypothesis-driven development. Alex explained the key is ITERATING, ITERATING, INTERATING! Alex also presented the topic modeling examples, including Reimagining the Monograph, which started from JSTOR Labs asking, "Can we improve the experience and value of long-form scholarship?"
The “topicgraph” provides a fingerprint of a monograph. Each term has a set of associated keywords, containment of which in the text make the probability higher that the term is being discussed.
Last but certainly not least, Alex unveiled am amazing and brand new tool with the working name “Text Analyzer.” This tool is essentially a multi-language analyzer—text can be pulled from, say, a Russian Wikipedia entry. The tool will translate the text and list in English the topics included in the entry.
Alex notes that so much of digital humanities is about probabilities, not known data. The label modelling that JSTOR Labs most frequently uses (as opposed to cluster topic modeling).