Monday, May 29, 2017

Dispatches from Cultural Analytics symposium

Visualization from our Jay Z Dataset presentation

By Kenton Rambsy and Howard Rambsy II

On May 26, we gave a presentation on our “Jay Z Dataset” as part of the Cultural Analytics symposium at Notre Dame. “Cultural analytics” refers to the use of computational methods and large datasets to identify and highlight patterns in production of various cultural artifacts. The symposium organizer Matthew Wilkens pulled together a range of scholars, who covered a range of topics broadly related to cultural analytics.   

Hoyt Long concentrated on racial markers based on “a corpus of several thousand fictional works from Japan’s colonial period.” Marit MacArthur presented her work on the performance styles of 100 American poets based on recordings of their poems. David Bamman and Richard Jean So presented their research about a computational model they developed to parse the story-based interactions between humans and objects within a large corpus of US novels published between 1880 and 2000. We talked about our findings by compiling a dataset of Jay Z’s 12 solo albums, which includes more than 175 songs and more than a hundred of Jay Z’s collaborators. The abstracts for all of the presentations are available here.

Some of the symposium presenters are in English, others are in Computer Science and Engineering, and still others are in Schools of Information Science. In some respects, the projects would be described as digital humanities, but several folks at the symposium actively question the limits of that term. Furthermore, DH doesn’t fully capture some of social science aspects of where people are and what they’re doing.

The symposium concluded with a roundtable – presenters and audience members literally sat in a circle – where we discussed opportunities and obstacles of the emerging field of cultural analytics alongside the broader field of digital humanities. The conversation covered a wide variety of topics, including data accessibility, ways to assist undergraduates with data analytics skillsets. Richard Jean So proposed that the humanities might shift towards being more problem oriented so that digital tools could be used to assist in testing the limits and possible solutions for various topics.

During the roundtable, symposium participant Ryan Heuser offered summaries of how far projects and approaches had developed in the field since previous symposiums on the topic in previous years. He noted that the presentations had collectively overcome some of its past challenges.

Speaking of challenges, Marit MacArthur addressed a current one: diversity. Symposiums in the realm of cultural analytics and DH still struggle to involve substantial numbers of women and non-white people. In addition to taking steps to diversify, we might not enough about why there’s a lack of diversity in the first place. So hopefully that’s something we can devote some of our energies to considering.

Underwood raised another important point. He noted that non-DH colleagues in the field will regularly identify DH projects that they find fascinating. What’s rare, however, is for non-DH colleagues to show interest in taking on those projects. An understanding of what prevents people from taking on DH or cultural analytics projects might a humanities problem worth further investigation.

The Jay Z Dataset--presentation at the University of Notre Dame

No comments: