NovelTM Graduate Students at the University of Notre Dame

Budget-1 notre dame

In May 2017, three NovelTM graduate students – Kaitlin Cruz, Melanie Walsh and Lisa Teichmann – met up in South Bend, Indiana, with Textual Geographies project director Matthew Wilkens for a week-long workshop dedicated to learning and exploring a range of geographical text analysis methods and tools, some of which are used in the NEH-funded project. At the end of the week, these graduate students also had the chance to participate in the annual Cultural Analytics Conference. Below is a brief account of the week from the graduate students’ perspectives.

Day I. Tutorial in geotagging and mapping in Python, organized by Matt Wilken’s

What did you learn?

Kaitlin: The first day was a lot of trial and error attempting to use Google’s geotagging (Cloud Services API) without it crashing. Using the Stanford “named entity recognition” package, we were able to retrieve location data from a text corpus. Through this process, we learned that it can be difficult to determine actual location due to places being named after people, as well as duplicate locations. Furthermore, the differences between countries/continents, states, cities, and streets presented a certain level of complication. However, the payout was worth the trouble once we created bubble maps of the mentioned locations, mapped onto a world map.

Melanie: Though I had used Python for text analysis before and had some experience working with APIs, I had not spent much time thinking about specifically geographical questions when it came to texts. Wielding this same basic skill set in a slightly different way, however–namely, adding Stanford NER and the GoogleMaps API into the mix – opened up a completely different and rich set of research possibilities. So while I did acquire new technical skills on the first day – and while I was particularly impressed and provoked by the prospect of using the GoogleMaps API, as well as the broader prospect of harnessing data and computing power from corporations – my biggest takeaway was simply learning that geographical questions could be explored within texts themselves.

Lisa: From reading and discussing Matt Wilkens piece “The Geographic Imagination of Civil War Era American Fiction” for one of my courses I was familiar with his approach to analyze geographical data in fiction. While being guided through using the Stanford NER and the GoogleMaps API during the workshop, the main question on my mind was: How does this apply to other languages and corpora of fiction? How does the NER as well as the GoogleMaps API perform on German versus English texts? In the months following the workshop I found myself going back to the material and code (you can find it here) we used and exploring these questions.

Day II. Meeting with the research staff and affiliates of the Textual Geographies project

What did you present on?

Kaitlin: At the TxtGeo meeting, I presented on a personal project. Drawing on a diary written by Gertrud Bleichröder, a well-to-do German-Jewish bourgeois woman, in Berlin in 1888 over a 6 month period, I compiled a list of mentioned names and places that she visited. The project was complicated by the fact that, although the lists of names and addresses of the individuals with whom she met had been digitized, the source font/script remained in an old German format. This proved impossible to OCR correctly, so instead I went through the digitized address/phone books manually to pull out the pertinent data. Thankfully, the resources were in alphabetical order! Once that was accomplished, spreadsheet in hand, I used ARC-GIS to map and track her movements over the period of time within which she wrote her diary.

Melanie: I presented work from my dissertation-in-progress, which examines the reimagination and recirculation of postwar American fiction by early 21st century readers, specifically about how quotations from the novelist and essayist James Baldwin have circulated within the Black Lives Matter movement on Twitter. Though I had not explored the geographical possibilities for Baldwin’s Twitter invocation–largely because the tweets from this time period (2014-2015) were, for the most part, not geotagged–I used this work to speculate about the broader possibilities of mapping text reuse and recirculation, not only on social media and not only in the 21st century. Tracking where and how far a particular word or phrase travels seems, to me, an extremely rich avenue of research.

Lisa: For the meeting I prepared a short summary of three of the DH/literary geography projects I have been working on in the past two years with a focus on the challenges and tasks:

    1. A Literary Map of Turkey: Mapping a corpus of Turkish texts (fiction and nonfiction in OpenStreeMaps)
    2. :aichinger: A mapping project with RDF annotations of geographical places in the complete works of the Austrian writer Ilse Aichinger (as part of the Aichinger research group at the Austrian Academy of Sciences).
    3. “Textual Networks of Fictional Space and Place”, an application of social network theories and methods (QAP) to spatial data extracted from the complete works of Ilse Aichinger. In this project I was especially interested in toponyms versus general denominators for places throughout the author’s work.

The meeting facilitated an environment to discuss my former projects and their possibly challenges with experts in the field as well as exchange ideas on general questions about the conceptual problematics in mapping and analyzing fictional texts spatially.

Day IV. Cultural Analytics conference

The conference program can be found here and a some videos taken during the conference here.

How do you think did the participation of grad students influence the discussion at the conference? What topics emerged?

Kaitlin: I feel that having graduate students participate in the discussion at the conference influenced the thought process of the presenters in a positive manner, broadening the scope and focus of topics and possible future research pathways. There seemed to be a lack of actual “mapping”, i.e. using cartography to display textual analysis results, despite the emphasis placed on textual mapping throughout the week of the Geospatial Workshop and the CA conference as well. Several themes emerged: the acknowledgement of a need for broader CA and digital textual analysis and computational analysis across all faculties; possible sensitive subjects, and why certain topics are being researched to begin with (for example, Hoyt Long and Yuancheng Zhu’s “Semantics of Race” presentation) which seemed more of a concern for the graduate students in attendance than the presenters themselves. This is something I believe needs more attention, why are we so focused on race, creed, gender, etc?

Melanie: Having graduate students in the room demonstrated, I think, a commitment to growing the field. It’s unique to have a small group of researchers who can share and engage with each other’s work at such a high level and with such intimate focus; inviting graduate students to be part of the conversation helps extend the expertise and insights fostered within this small community as well as provides outside perspectives and challenges from different institutional positions.

Lisa: The meeting conveyed a wide array of current research in and beyond digital humanities. I found the discussions around data ethics and the role of digital humanities research on developments regarding open access very interesting. Especially … presentation raised points that lead to a lively discussion. … presentation also drew attention to digital humanities tools in the classroom.

What did you take home from your week at Notre Dame?

Kaitlin: After a week at Notre Dame, I felt recharged and had a new-found interest in DH. I wish there was more inclusion across the board between professors and graduate students regarding DH training and projects, which I feel would benefit both sides. Having so many amazing resources available can be daunting, and since practice makes perfect, I feel that holding more of these workshops would be beneficial to the field in the future.

Melanie: I similarly came away thinking that the workshop model is an excellent way for graduate students to be exposed to different computational skills and methods – if not hone them completely – and an excellent way for graduate students to get to know each other and to form collaborative collegial relationships. I was also struck by the value of small conferences, specifically in terms of how deep and engaged the feedback can be and how quickly a sense of community can be forged.

Lisa: Besides the academic benefits of student workshops hosted by a partner University of NovelTM, the time the students spent together also turned out to form the basis for future collaboration in common areas of research.

Last but not least, are there any resources (tools, articles, etc.) – including your own work– you would like to share, that you discovered through the workshop?

Melanie: I became really interested in word vectors after our week at Notre Dame, and I’ve found Ryan Heuser’s and Teddy Roland’s code and tutorials especially interesting and useful.

The participants want to express their gratitude and thanks to Matt Wilkens and the University of Notre Dame for hosting this workshop.