2025 draft

AI for Accessible Justice: An Affordable and Transparent Legal Solution

Peter Nadel, Jill Weinberg (Tufts University)

This paper presents results of research assessing the efficacy of a novel, transparent LLM designed to analyze legal contracts, specifically social media Terms of Use agreements. Using a large language model (LLM) trained on legal materials, this research incorporates unique interpretability tools that explicitly demonstrate its decision-making process to end users. The research also involved rigorously testing the model's robustness to decipher errors, such as typographical errors, negation manipulations, and legally incorrect content, mimicking a role of an uninformed legal user (i.e., a pro se litigant). In addition to offering a clear assessment of compliance or non-compliance, the research also focuses on precise error identification with user-friendly explanations, qualitative analysis of the AI's decision-making language, and systematic evaluation of potential biases. By advancing the field of explainable AI and promoting responsible implementation, this research contributes to the broader goal of increasing access to justice. This paper also presents a framework for developing AI legal tools that balance effectiveness with user trust through systematic transparency. Findings suggest that transparent AI can significantly improve user trust and accessibility to legal information, particularly for underserved communities.

Jill Weinberg is an Associate Professor of Sociology at Tufts University whose interests are research methods and socio-legal studies. Peter Nadel is a member of the Research Technology Team at Tufts University and is a digital humanities and natural language processing specialist.

The World Salvage Project: Interdisciplinary Futures and Possibilities for Digital Repositories in the Age of AI

Sybil F. Joslyn (Boston University)

As Professor of the Humanities William Condee noted in 2016, "the arts and humanities are in the midst of an interdisciplinary turn." Coinciding with a nearly 25% drop in awarded humanities degrees in the United States from 2012-2020, this interdisciplinary turn reflects an instinctual need for humanistic disciplines to remain relevant as the academy itself shakes with change. As humanistic disciplines forge new identities, Digital Humanities initiatives should be at the forefront of embracing interdisciplinarity, modernizing humanistic research, and embracing innovative technological tools. In 2024, I created the World Salvage Project (WSP) as a digital data repository to accompany my dissertation on maritime salvage in America's long nineteenth century. Identifying its usefulness to other scholars, I published the WSP as a database-in-progress that unites individual ship records with other associated extant media, from objects and renderings to archaeological records. In my talk, I will introduce the origins of the project as an interdisciplinary repository, the benefits of utilizing Digital Humanities projects as dissertation research tools, and the challenges I have encountered in implementing best practices with limited resources. I will then discuss possible futures for the WSP, including how its development can be served using Artificial Intelligence agents, the importance of open-access and institutional collaboration, and the options that exist to safeguard its data. Through discussing the WSP, I hope to highlight how Digital Humanities initiatives can attract diverse audiences, connect scholars and institutions worldwide, and create a future in which shared knowledge across disciplines drives and aids humanistic research.

Sybil F. Joslyn is a fifth-year PhD Candidate at Boston University studying American art and material culture in the long nineteenth century. Grounded in object analysis and interdisciplinarity, her research explores the junction between material and visual culture, the social and economic life of objects, and intercultural exchange in the Atlantic World. Sybil's dissertation, "Worth Its Salt: Salvage in the Maritime Visual and Material Culture of America's Long Nineteenth Century," examines maritime salvage as object, material, and process to interrogate conceptions of identity, property, and value during the Age of Sail.

Broadcast News Summarization with Large Language Models

Kelley Lynch (Brandeis University)

This talk proposes a modular pipeline for summarizing broadcast news videos, leveraging the power of large language and vision models. We will detail the integration of Whisper for ASR, TransNetV2 for shot segmentation, LLaVA for image captioning, and LLaMA for generating structured summaries within the CLAMS open-source AI-assisted metadata extraction platform. A key focus will be demonstrating how combining ASR transcripts with shot-level image captions enhances metadata extraction, effectively addressing challenges such as ASR errors and enabling the generation of more informative summaries. We will present a comprehensive evaluation of the pipeline using both automated metrics and human assessments, benchmarked against user-generated YouTube descriptions. This evaluation will offer insights into the strengths and limitations of each evaluation approach.

Kelley is a Ph.D. candidate at Brandeis University, specializing in natural language processing and computer vision. As a developer on the CLAMS (Computational Linguistics Applications for Multimedia Services) Project, she focuses on creating tools for information extraction and summarization of archival broadcast news media.

Digital Engagement with Art Collections

Peter Botticelli, Heather Hole (Simmons University School of Library and Information Science)

We will discuss an early-stage project in which we are exploring ways to expand the options for scholarly engagement with art collections, in both material and digital form. Our initial focus involves working with the Isabella Stewart Gardner Museum to investigate the use of digital surrogates for both collection objects and archival records, with the aim to expose hidden contexts and to facilitate new modes of storytelling around the museum. We will present a number of case examples of collection objects and related records in the museum archives, and we will discuss the use of the collection in courses we have developed at Simmons. A larger aim of this work is to examine how digital surrogates can be used to augment as well as to distort the context in which museum objects may be viewed, with the aim to understand how museums might develop or update their access policies to account for AI applications.

Peter Botticelli is an Associate Professor at the Simmons University School of Library and Information Science (SLIS). His teaching and research are focused on archives management, and he directs the interdisciplinary Cultural Heritage Informatics concentration. Botticelli has A.M. and Ph.D. degrees in History from the University of Illinois at Urbana-Champaign, and an M.S.I. degree with a concentration in Archives and Records Management from the University of Michigan. Heather Hole is a Professor of Museum Studies and Art History at Simmons University. Her scholarship examines the history of American art and museums in the twentieth century. A former curator at MFA Boston and the Georgia O’Keeffe Museum, Professor Hole received her Ph.D. in Art History from Princeton University and her B.A. in Art History from Smith College.

Turning Almanacs into Data

Ashlyn Stewart, Joanna Schroeder, Yuchen Xiong (Tufts University)

The Boston College Libraries Catholic Collections Initiative has launched a multi-year, interdisciplinary project to extract, structure, and make available data about Catholic institutions from a series of nineteenth-century Catholic almanacs. Each almanac includes a list of all the Catholic institutions in each of the country's dioceses from that year—every school, hospital, orphanage, asylum and more–including the institution's name, location, and leaders. Our presentation will include three parts to explain our project in progress, highlighting the choices we have made to improve our dataset, our project workflows, and the experience of our users. First, Ashlyn will discuss a few key decisions she has had to make about regularizing heterogeneous data from this quirky source. She will focus on breaking the data into subsets and associating and disambiguating names of individuals in the almanacs. Next, Joanna will explain how she has improved the project’s data management with automation. Because the project runs on GoogleSheets, she has used Google Apps Script and the Sheets API to aggregate and standardize data across the project. For example, she built a system that updates on a daily basis to assign unique IDs to Catholic leaders. Finally, Yuchen will demonstrate her approach to the interconnectedness of the data and her process of making it searchable. She will illustrate how she has preprocessed the metadata on GoogleSheets with python, built a database with SQL, and created a web interface with Django in order to help web users find specific institutions and individuals.

*We are submitting two proposals focusing on different parts of this same project. We would be happy to expand our discussion to fill all or part of a panel, if that helps, or combine our efforts in some way.

Ashlyn Stewart is a Digital Scholarship Specialist in the Boston College Libraries Digital Scholarship Group, where she collaborates with students and faculty to create and maintain digital projects inside and beyond the classroom. One of her main responsibilities is leading the BC’s Catholic Almanacs project. She holds a PhD in English from the University of Nebraska-Lincoln, where she worked at the Center for Digital Research in the Humanities, contributing to the Walt Whitman Archive and the Charles W. Chesnutt Archive. Her research interests include nineteenth-century American literature, history, and culture, as well as digital archives and editions.

Joanna Schroeder applies her expertise in finding and using data in creative ways in her role as Data Services Librarian in Boston College’s Digital Scholarship Group. Previously, she worked as a data science research specialist at the University of Virginia’s Biocomplexity Institute, where she used data analysis and visualization skills to provide local, actionable metrics on topics like food security, broadband availability, and access to health services. She received her MLIS from Drexel University.

Yuchen Xiong is a first-year English MA student at Boston College, primarily working on the website development of BC’s Catholic Almanacs project. She holds a Bachelor of Engineering degree in Automation from Tsinghua University. With a multidisciplinary background, Yuchen is passionate about exploring the intersection of humanities and technology. Her research interests focus on how Digital Humanities methodologies, such as database and textual analysis, can bring new insights into traditional literary studies.

Building a Digital Trade Archive: Visualizing British East India Company Data (1660–1834)

Munirmahedi Paviwala (Boston College)

This presentation examines the challenges of integrating historical trade data on the British East India Company (EIC) and how visualization can aid historical inquiry. I use datasets from the works of K. N. Chaudhuri and H. V. Bowen which covers the EIC trade from 1660 to 1834. I highlight how the choices made in cleaning and merging can shape the final visualization.

The talk will outline key decisions made in data integration, including how to deal with missing or fragmented records. I visualise the merged data using Tableau. The mapping of the trade flows reveals patterns and prompts new research questions. The project is open-access, with all data available via GitHub for further collaboration.

This presentation aligns with the symposium’s themes of public engagement and digital methods in humanities research, showing how historical datasets can serve as a foundation for deeper inquiry. Instead of a singular narrative, my project provides a visualisation that allows historians to study EIC trade networks for their own questions.

Munirmahedi Paviwala is a historian and digital scholar passionate about making historical research accessible. Munir is currently a PhD student at Boston College, his work focuses on environmental and economic histories of South Asia. He employs digital humanities tools to create interactive and publicly engaging historical narratives. His projects have visualized British colonial trade networks and agricultural transformations in India. Munir holds a BA (Hons.) in Economics and a Postgraduate Diploma in History from Krea University, India. He also earned a Certificate in Digital Humanities from Boston College.

The Ground Beneath Our Feet: Critical Mapping in the Classroom

Kelley Kreitz (Pace University); Anindita Basu Sempere, Jamie Folsom (Performant Software)

The Ground Beneath Our Feet is a place-based, experiential humanities research and curriculum initiative at Pace University that was launched in Fall 2024, with support from a NEH Humanities Initiatives grant and the university’s Student Technology Fee Fund. Students in affiliated courses research and document the history and the communities surrounding Pace campuses, using critical mapping to juxtapose their contemporary, firsthand experiences of Lower Manhattan and the Hudson Valley with their place-centered research. Course topics include New York City's 19th Century Spanish-Language Press, Hurricane Sandy, Counter-Mapping Governance and Community, and Cultural Cartography, and the data and project maps can be related and/or layered to show intersections and differences between datasets. Through custom-built, open source tools, the Ground Beneath Our Feet platform provides support for different types of place-centered storytelling: Places (a searchable map), Posts (essays with embedded places), and Paths (curated sequences of places).

This talk will describe the overall vision of this project, the supporting technologies, the initial round of critical mapping in the classroom, and next steps based on student and instructor feedback. As we discuss the pedagogical importance of providing multiple tools for combining data and narrative in studies of place, we will also share our process, realizations, successes, and challenges as we seek to empower students to put technology to work for what Ruja Benjamin calls “retooling solidarity, reimagining social justice.”[1] We look forward to sharing ideas and experiences with participants.

[1] Ruja Benjamin, Race After Technology, (Cambridge, UK: Polity Press, 2019), 160.

Kelley Kreitz is Associate Professor of English and Interim Associate Provost of Academic Programs and Strategy at Pace University in New York City. Her research brings together Latinx Studies, U.S. literary studies, media studies, and the digital humanities. Her book, Printing Nueva York: Spanish-Language Print Culture, Media Change, and Democracy in the Nineteenth Century, is forthcoming from NYU Press. She was the Co-Director of a 2021 NEH working group on digital mapping in Latinx studies and serves on the advisory boards of the Recovering the US Hispanic Literary Heritage Project and New York City Digital Humanities (NYCDH).

Anindita Basu Sempere is a Project Manager at Performant Software, where she works with scholars and software developers to design and deliver digital humanities projects. She has a PhD from the University of Neuchâtel’s Institute of English Studies, where she studied change of place and poetics. As part of her doctoral research, she combined digital humanities methods with genetic criticism to study the intersection of place and literary creation. She holds Masters degrees from the MIT Media Lab’s Future of Learning Group, Boston University’s Poetry Program, and the Vermont College of Fine Arts.

Jamie Folsom is a Partner at Performant Software Solutions, an established software development agency with a focus on the Digital Humanities. He consults with scholars in a wide range of fields, managing technology projects with them, and collaboratively developing software tools to support their teaching and research. Previously, he was lead developer for Hyperstudio, center for the Digital Humanities at MIT, where he worked on projects including Annotation Studio and the Comédie Française Registers Project, among others. He holds a Bachelor's degree in French from Vassar College and a Master's degree in Technology, Innovation, and Education from Harvard University Graduate School of Education.

How the Rhetorical Choices of Syllabi Can Enact Black DH Pedagogy

Laura Joy Clarke (Northeastern University)

The recent publication of books such as Doing Black Digital Humanities with Radical Intentionality by Steele et al. signals a growing concern with teaching DH in thoughtful, student-empowering ways. One easily-overlooked element of pedagogical design is the rhetoric used in syllabi. My research addresses this gap by taking a deep dive into a corpus of DH syllabi: Specifically, I examine their structure and components, as well as sentence-level rhetorical choices to determine how “student-oriented” versus “instructor-oriented” or “content-oriented” these documents are. I unpack how students are discursively constructed in these texts through the nouns and verbs associated with them and discuss the bearing these constructions may have on their perceptions of their own abilities and agency. My corpus contains two sets of syllabi: three from generic “Intro to DH” courses and three from race-focused or Black DH courses. Across all six syllabi, I found examples of the myriad ways that Black DH pedagogies can be enacted (or not) through seemingly simple, even sentence-level decisions. My talk concludes with practical tips and considerations for instructors to implement in their own syllabi. After all, the way we compose our syllabi offers a unique opportunity to begin empowering students before they even set foot in the classroom.

Laura Joy Clarke is pursuing a master’s degree in English as well as a certificate in the Digital Humanities at Northeastern University. Her research interests include composition pedagogy, rhetorical analysis, and educational policy. In her spare time, she also moonlights as a technical writer, a writing tutor, and a book editor.

Incorporating Undergraduate Students into Large DH Projects

Kaitlyn Bell, Nathanael Choi, Ashlyn Stewart (Boston College)

At the Boston College Libraries, we are fortunate to have twelve undergraduate and two graduate students working on one of the Digital Scholarship Group’s flagship projects (in conjunction with the Catholic Collections Initiative) this semester. These paid student workers spend anywhere from five to fifteen hours a week extracting, structuring, and manipulating messy data from a series of nineteenth-century Catholic almanacs. While much of the project is collaborative, we have attempted to break the undergraduates’ work into localized portions so that one student’s progress does not immediately hinge upon another student’s. This project has tested our project management pipelines, allowing us to learn more about how to manage a dispersed, asynchronous team where almost all of the data creation is done by undergraduate student workers. Ashlyn will discuss how she incorporates students into the project to suit their individual learning, interests, and strengths. She will also demonstrate how she oversees and checks their work using mostly GoogleDrive. Two undergraduate members of the team will join her to share how they have contributed to the project and what they have learned about working on such a large research initiative. Kaitlyn will discuss her work of improving location information for thousands of institutions, demonstrating the most micro of tasks that our students complete. At the macro end of the scale, Nathanael will share his efforts in writing and implementing Google Apps Script code across all of the project’s GoogleSheets at once.

*We are submitting two proposals focusing on different parts of this same project. We would be happy to expand our discussion to fill all or part of a panel, if that helps, or combine our efforts in some way.

Nathanael Choi is a junior at Boston College originally from Long Beach, California. He is majoring in Mathematics with minors in Data Science and Business. He works as a student assistant for the BC Libraries Digital Scholarship Group, applying his technical skills to the Catholic Almanacs project to analyze dense historical data and uncover compelling stories. Though focused on numbers in his studies, Nathanael is passionate about the humanities, especially literary fiction; his favorite novel is Never Let Me Go by Kazuo Ishiguro. He’s also involved in Club Ping Pong, the AHANA+ Pre-Law Student Association, and Campus Ministry.

Kaitlyn Bell is a sophomore at Boston College from Burlington, Connecticut. She is a major in History with a minor in English and is specifically interested in the medieval and early modern worlds. She works as a Research Assistant on the Catholics Almanacs project, while also working as an Administrative Assistant and the Instagram Manager for the BC History Department. Next year she will serve as the Resident Assistant for the Women’s Experience Living-Learning Community for freshmen. In her free time, she enjoys spending time with her family (including poodle Sweetie), reading, playing golf, and watching sports.

Ashlyn Stewart is a Digital Scholarship Specialist in the Boston College Libraries Digital Scholarship Group, where she collaborates with students and faculty to create and maintain digital projects inside and beyond the classroom. One of her main responsibilities is leading BC’s Catholic Almanacs project. She holds a PhD in English from the University of Nebraska-Lincoln, where she worked at the Center for Digital Research in the Humanities, contributing to the Walt Whitman Archive and the Charles W. Chesnutt Archive. Her research interests include nineteenth-century American literature, history, and culture, as well as digital archives and editions.

Public knowledge in crisis: Mining Wikipedia's digital history of the opioid epidemic

Rona Aviram (Weizmann Institute of Science, Israel; Learning Planet Institute, Université Paris Cité, France)

Wikipedia, founded to overcome the limitations of print-based encyclopedias through digital tools like hypertext and open editing, has become a dominant platform for public knowledge-sharing. Previous works by us and others, show it can also serve as a powerful resource for historical research. Our methodology combines traditional and digital humanities approaches, performing systematic diachronic analyses of Wikipedia articles’ text, edit histories, hyperlinks, and references to understand their evolution over time. Here, we applied these methods to study English Wikipedia's portrayal of the opioid epidemic, a critical and global public health crisis. A custom-built “corpus builder” tool identified 241 relevant Wikipedia articles, extracted their metadata including past edits, contributors, and bibliometric data. This information was qualitatively contextualized to trace the development of key narratives. Our findings reveal a delayed response in Wikipedia’s coverage of the opioid crisis, with key articles appearing well after the crisis gained public and academic attention. Moreover, treatment discussions primarily focused on drug-based approaches, underplaying the risks of opioid addiction, reflecting a distressing reality. Examining the connections between articles in present and past forms, we found significant knowledge communication failures in the coverage of opioid-induced hyperalgesia, an underrepresented aspect of opioid misuse. Our findings highlight gaps in Wikipedia’s coverage of the opioid crisis that reflect broader deficiencies in public discourse and medical policy responses. More broadly, it demonstrates how Wikipedia can be used as a living, digital archive for examining evolving public discourse, and offers reproducible tools for other researchers to capture the incremental growth of knowledge.

As an early career researcher working at the intersection of digital humanities, history of science and computer science, I explore how scientific knowledge is organized and disseminated digitally. My PhD training was as a wet lab biologist, examining the molecular basis of daily rhythms and sleep. Currently, as a postdoc at the Weizmann Institute's Dept. of Computer Science and Applied Mathematics, I am developing computational tools and methodologies to study the broader processes of science-making and knowledge production, and the manner scientific discoveries are reflected in online digital archives, such as Wikipedia.

Salus Populi Project: Digital Archives, Reparative Genealogy, and Restorative Justice

Riley Sutherland (Salus Populi Project; Harvard University)

The Salus Populi Project is a non-profit digital humanities project based in Liberty, Missouri. We are currently creating a publicly accessible online archive for the pensions of United States Colored Troops veterans who were born or enslaved in the counties of Missouri colloquially known as “Little Dixie.” White residents of these counties enslaved one-third of Missouri’s unfree population, many of whom liberated themselves to serve in the Union Army. After the war, they applied for pensions, and their files contain paperwork related to freedom seekers’ experiences of enslavement, military service, families, disabilities, and community networks. But none of the files are digitized, transcribed, or searchable.

The Salus Populi Project is partnering with local archives, historical societies, libraries, museums, and descendant communities to digitize, transcribe, and contextualize these records. We are utilizing From the Page, a platform with crowdsourcing capacities, to create a more participatory project. We are also utilizing digital mapping tools to visualize soldiers’ journeys and networks, and partnering with students to narrate stories through short digital films.

This talk will address the challenges and possibilities we have encountered while creating a digital archive that offers space for descendants, educators, and students to grapple with traumatic histories of racism and grapple with slavery’s legacies.

Riley K. Sutherland (she/her) is a History Ph.D. student at Harvard University, where she specializes in nineteenth-century U.S. history, disability studies, and public history. She worked as an editorial assistant to the Pinckney Papers Project and as an archival intern for the Harry S. Truman Presidential Library and Museum. She currently co-directs the Salus Populi Project with historian and artist Michelle Cook.

The Design and Technical Development of a Version Update: The Burnham-Nobles Digital Archive

Candace Hazlett, Joel Lee (Northeastern University Library)

The Digital Scholarship Group (DSG) at Northeastern University Library provides expertise in digital infrastructure and technical design to the Civil Rights and Restorative Justice (CRRJ) Project to create the Burnham-Nobles Digital Archive (Archive), a digital resource dedicated to identifying, classifying, and providing factual information and documentation about anti-Black killings in the mid-twentieth century South.

As the Archive nears its first update, which expands its geographical scope and includes new cases, we, as members of the DSG, share the technical process that supported these cases as they were collected, researched, and added to the Archive and web application. We touch upon how we realize the goals of the project through use of a variety of different tools like Airtable and Clio for the various steps of the research, Python for processing the records into a tabular format, as well as PostgreSQL, Ruby, and Solr to bring that research into the Archive web application. Our presentation will be focused on outlining this technical scaffolding and sharing the intentional technical design and development decisions, from tool selection to the collaborative design process for the Archive UI/UX.

The Archive seeks to render these difficult histories as visible and accessible, so as to honor the victims and their families and support the efforts for redress and accountability. Towards that end, we also touch upon the technical and process steps of our work in order to mitigate dehumanization and honor the victims while building a database that contains harmful materials.

Candace Hazlett is a software developer for the Digital Scholarship Group at Northeastern University’s Snell Library. Currently, she is developing the Burnham-Nobles Digital Archive for the Civil Rights Restorative Justice project, along with a revamp of the TAPAS web application for publishing TEI documents. Candace is a Massachusetts native and prior to joining DSG, developed software for local healthcare tech companies. She spends her free time reading, coding, watching Netflix and YouTube, spending time with her family, volunteering, and glamping when weather permits.

Joel Lee is a data engineer for the Digital Scholarship Group at Northeastern University's Snell Library. He works on the Boston Research Center, the Civil Rights and Restorative Justice Project, and projects with Digital Humanities Quarterly and the Massachusetts Historical Society. He was the 2021-2022 Digital Humanities Associate Fellow at the United States Holocaust Memorial Museum and before that he worked with the Price Lab for Digital Humanities while studying at the University of Pennsylvania.

Gone in an Adobe Flash: Five new frameworks to preserve born-computational literary art for the future

Ashley Champagne, Khanh Vo (Brown University Library)

Brown University is the source of some of the earliest, most experimental and innovative works in the field of born-digital literary arts. Faculty and students at Brown regularly produce work in a variety of experimental formats designed for digital environments, from neural-net and model-based aesthetic text-generation programs to fully immersive three-dimensional poetry, and have done so since the early 1960s. Like many institutions, however, Brown University does not yet have models or strategies to preserve many of these works that push boundaries of electronic literature. This talk provides a case study of the university’s New Frameworks to Preserve and Publish Born-Digital Art project, which approaches preservation by focusing (1) on the works that are difficult to preserve in order to determine what approaches will work best for them, and (2) to encourage artists working in these areas to pursue technically innovative art with the knowledge of how to go about sustaining their work for future generations. The talk outlines the preservation frameworks developed for five innovative works that rely on dependencies including Adobe Flash (a technology that recently reached its end of life), aging libraries, browser extensions and broken custom code. The frameworks outlined aim to provide a sustainable future for accessing these five works, as well as a generalized guide for institutions and artists alike in how to preserve difficult, technologically innovative art for future generations. We will also share Brown University Library’s future plans to preserve born-computational work.

Dr. Ashley Champagne is the Director of the Center for Digital Scholarship, Brown University Library. She is the PI of “New Frameworks to Preserve and Publish Born-Digital Art,” a project funded by the National Endowment for the Arts. She earned her Ph.D. in English from University of California, Santa Barbara. Her work has been featured in the Brown Alumni Magazine (2023), IMPACT Research at Brown Magazine (2023), Interdisciplinary Digital Engagement in Arts & Humanities (IDEAH) (2023), and the Journal of Digital Media Management (2025).

Dr. Khanh Vo is the Digital Humanities Specialist at CDS. Her background is in American Studies, specializing in American history, STS, the digital and spatial humanities, and machine labor as it pertains to questions of race, gender, and design. Prior to Brown, she was the Postdoctoral Fellow at the University of Toronto’s Jackman Humanities Institute and Critical Digital Humanities. She has a PhD in American Studies from William & Mary.

“A Scientific Association”: New Digital Methods for Understanding the Impacts of Early Women Writers on the Development of Science and Philosophy

Melissa Williams (Northeastern University)

This paper will share insights from a new initiative studying the impacts of women on the sciences during the watershed period of the seventeenth century in Great Britain. The project, led jointly at Northeastern’s campuses in Boston and London, is using text and network analysis to investigate the ways that women scientists and natural philosophers such as Margaret Cavendish shaped emerging scientific and philosophical discourses. In partnership with the PolyGraphs project, we are performing network analysis to study the paths through which these authors circulated their works and ideas, especially in relationship with the networks of cornerstone scientific institutions like the Royal Society. The project is also collaborating with the Women Writers Project to expand the corpus of women’s scientific writing and to analyze this corpus through TEI/XML, topic modeling, word vector modeling, and other methods.

Preliminary results have already demonstrated the breadth and significance of women writers such as Cavendish, whose reputation for eccentricity has too often overshadowed her substantive engagements with the science of her time, as is made clear through the range of her scientific publications, the extent of her citation networks, and the saturation of scientific discourse not just in her nonfiction but throughout her poetry and plays. This paper will share outcomes from the project’s efforts to enable novel research into this seminal period and to provide new tools for raising the public visibility of women and the sciences, helping to show that women have contributed to scientific and philosophical disciplines since before the beginnings of modernity.

Melissa (Mel) Williams is a first year PhD student in English at Northeastern University. Her fields of study primarily include colonial and post-colonial theory, multiculturalism, and the role of literature in education. Additionally, she assists with several Digital Humanities projects such as the Margaret Cavendish project where she serves as graduate student lead for the NEU Boston team.

Hacking the Archive: Digital Archiving, Activism, and the Future of Public Memory

Shireen Zaineb (MIT)

Engaging with the MIT-based project Hacking the Archive as a case study, this session examines the role of community-driven research and digital technology in creating post-custodial archives. The project aims to build a digital archive that not only highlights the individuals and communities involved in Boston-area social justice actions such as the 1960s anti-highway fight, but also displays the connections between them. In doing so, the session explores three key questions: How can an archive highlighting community connections help us understand and theorize social movement formations? How can a born-digital, networked representation of activism position itself to be accessible to all members of the public? In what ways can an archive of community activists challenge our understanding of time and public memory while also informing future civic actions? In order to address these questions, the presentation explores the methods and strategies developed by Hacking the Archive’s team through cross-institutional and community-oriented collaboration. The project ultimately considers and engages the challenges of establishing a digital, open access archive while maintaining the necessity of publicly available records that preserve and sustain collective memory.

Shireen Zaineb is a research assistant at MIT’s Hacking the Archive project. Having recently completed her master’s degree in Islamic Studies and Muslim Cultures at Columbia University, her work focuses on the intersection of technology, community collaboration, and liberatory archival practices. Drawing on her background in media studies, multilingual storytelling, and cultural preservation, she is dedicated to preserving histories and uplifting archives that empower collective memory and imagination.

Computer Vision for Distant Viewing of Video Materials

Owen C. King (GBH Archives); Kyeongmin Rim, Marc Verhagen (Brandeis University)

Imagine you’re a researcher who has just gained access to the personal manuscripts of a significant but under-studied political leader. But suppose that, unfortunately, all the boxes and folders are unlabeled, and there is no finding aid for the collection. How might you even begin? Media studies scholars face an analogous predicament when attempting to engage collections of digitized videos. Because analog media deteriorates, digitization efforts for old video tapes are often undertaken under conditions of relative ignorance about the content on those tapes. The result is digital video files, perhaps with dates and/or titles, but little indication of what the footage actually encompasses. One could simply watch all the videos, but that is impractical for large collections. But manually scrubbing through videos is imprecise and error-prone. Our solution is a system for automatically creating visual indexes of digital videos, for identification and display of scenes with text that identify the persons and places featured in the videos. The visual indexes are based on a computer vision model we trained to classify scene types, especially scenes with text, from broadcast television, for further processing with language understanding models to pull out personal names. This presentation explains our development and distribution of a system for extracting information and creating visual indexes. This includes: (1) our specialized training data, (2) training and evaluating our computer vision model, (3) the multimedia interchange format (MMIF) for storing media annotations, (4) the design of visual indexes to support a variety of cataloging and analysis use cases, (5) distribution of the models-and-visualization pipeline as an open-source software package. The final section of the presentation will be a demonstration of how to acquire and use the full package to create visual indexes for your own digital videos.

Our response to conference committee: As we revised and re-submitting this proposal, we recognize that our original abstract was geared more toward archivists than to researchers. We think that the same techniques are broadly applicable to both. We have reframed our abstract accordingly. If this topic still falls outside the scope of this conference, then we understand.

Owen C. King is Metadata Operations Specialist at GBH Archives, working on the American Archive of Public Broadcasting. He coordinates work on automated processing and metadata workflows for archival television and radio programming. He has a PhD in philosophy from The Ohio State University and an MS in information science from the University of North Carolina at Chapel Hill.

Kyeongmin Rim is a PhD student working on computational semantics. After working at an industry-leading NLP company in Korea, he returned to academia pursuing Computational Linguistics at Brandeis University. As a doctoral research assistant, he’s been working on various projects, building multimodal and multimedia semantic models for artificial intelligence systems. He is lead software developer of the CLAMS platform.

Marc Verhagen is a Senior Research Scientist at the Brandeis University Computer Science Department. He holds a BA in Geography and an MA and PhD in Computational Linguistics. He was co-founder of Lingomotors, where he acted as Director of Tool Development and designed and co-implemented the company's Knowledge Management tools. At Brandeis he has been involved in temporal awareness, advanced information retrieval and the development of CLAMS, a platform for integrating processing applications for multi-modal data.

Building a Platform to host 3D models

David Thomas (Boston College)

The Boston College Libraries Digital Scholarship Group is developing a lightweight, open-source solution for hosting 3D models. The recent uncertainty surrounding Sketchfab's commitment to hosting open-access models has posed significant challenges for universities, libraries, galleries, and museums reliant on such services. In response, our team is creating a user-friendly platform that allows institutions and individual researchers to host 3D models efficiently and affordably, tentatively called RelicQuery. This solution is designed for flexibility: smaller collections can be hosted on GitHub Pages, while larger collections can be self-hosted. Users simply upload a folder of models alongside a spreadsheet containing relevant metadata. This approach not only empowers institutions to preserve and share their collections but also enables student research projects to host 3D models without straining university resources. In this presentation, our developer will demonstrate the platform, explain its functionality, and invite feedback from the community to inform future enhancements. This project represents an exciting step forward in ensuring sustainable, accessible infrastructure for 3D content in the digital humanities.

David Thomas is a digital scholarship specialist in the Boston College Libraries Digital Scholarship Group who works at the intersection of ancient history and digital humanities, as well as in digital scholarship more broadly. He holds an MA in history from Brown University and Northern Illinois University. He was the sole developer of the Networks of Roman Eleusis project and has published packages in Python. He has worked in text analysis, network analysis, and more, but most of all, he is a full-stack web programmer who helps students and faculty build digital projects for the web in a sustainable way.

Digitizing and Exposing the Cambridge Survey of Architectural History

Meta Partenheimer (Cambridge Historical Commission), Paul Cote (pbc Geographic Information Services)

Meta and Paul will present the Cambridge Digital Architectural Survey and History (CDASH), comprising approximately 2,500 folders, 140,000 pages, 60,000 documents concerning approximately 20,000 Cambridge buildings and other places. They will introduce how the 36 file drawers of folders and documents were prepared for scanning, and the naming convention that allowed the scanned collection to be digested into a hierarchical network of Place, Folder, and Document items in Omeka-S. The tools and workflows for semi-automated document sorting and categorization, metadata enrichment and georeferencing, and bulk upload to Omeka will be discussed. Paul will talk about the CDASH theme for Omeka that provides a web-map interface that places the collection in the context of contemporary and antique old maps shared by Cambridge GIS, and other cultural institutions.

Meta Partenheimer is the Digital Projects Archivist for the City of Cambridge Historical Commission (CHC) and Project Manager for the Cambridge Digital Architectural Survey and History (CDASH). Meta has been working in cultural institutions for over a decade and received her MLIS from Simmons College (now University) in 2018. In addition to the CDASH project, Meta participates in all aspects of archival work from donor relations to collections care, digitization, reference and outreach, and management of CHC social media.; Paul Cote has worked on geographical projects around Boston and Cambridge since 1993. He has been an instructor at Harvard Graduate School of Design, Tufts and the Boston Architectural College. Since 2013, Paul has been the principal consultant for pbcGIS.com, working with the GIS departments of Boston and Cambridge on archival systems for preserving and sharing 3D city models. Visit https://pbcGIS.com for more information about Paul and his projects.

Digital Tools for Ancient Language Research

Alicia Tu (Tufts University)

Ancient language research is much more restricted than modern language research, since there are fewer translators fluent in the ancient languages, and the texts are oftentimes in bad condition. Applying digital tools to ancient language research can help with these problems. With Optical Character Recognition (OCR) tools, texts on fragile copies of books, large monuments, and other materials hard to access and work with can be converted to editable texts in digital forms, making research more convenient while causing less damage to the physical texts when studying them. Large Language Models (LLMs) show high accuracy in translating Classical languages into English, proving useful in providing a basic translation to assist scholars. Finally, digitization and translation should be available online, open for more scholars who do not have direct access to the texts.

Research done on the Latin and Classical Chinese translations of the Xi’an Stele, accessed from Anthanasius Kircher’s “China Illustrata,” is used as the case study to demonstrate the application of digital tools. Erected in 781 CE, this stele recorded Alopen, the first Assyrian missionary, arriving in China in 635 CE, bringing sacred books and building the first Christian church with other Syriac missionaries with the consent of the Tang dynasty emperor. This research project provides full text online in Latin and Classical Chinese, as well as English translations, compares the accuracy of Kircher’s Latin translations to the original, and spots interesting expressions used when Chinese writers and Western writers adapt to each other’s cultural concepts.

Alicia Tu graduated from Brandeis University with a BS in Computer Science, a BA in Classics, and a BA in Linguistics in 2023, and is currently pursuing a MA in Digital Tools for Premodern Studies at Tufts University. Her academic interest lies in studying ancient languages (mainly Ancient Greek, Latin, and Classical Chinese) with digital tools and methods. She also works part-time at Special Collections at Tisch Library as a student assistant, and seeks to incorporate more digital methods onto Special Collection books.

Tadhg Dall Ó hUiginn: Is é seo mo bhaile (The Ecos of Tadhg Dall Ó hUiginn)- A Decolonial DH Deep Map

Melissa Buckheit (Boston College)

The bardic poet, Tadhg Dall O hUiginn (1550-1591), of Leyney, Mayo, Connaught, was one the most prolific poets of the Gaelic literary tradition in early modern Ireland. Writing in the midst of the Tudor conquest in Ireland–a period of extreme violence, social and cultural alteration, dispossession and linguistic colonization–as the bardic poet for a far–reaching range of Gaelic lords and patrons throughout Ulster and Connaught–from Cú Chonnacht Mág Uidhir to Aodh Mág Uidhir, Brian na Múrtha, Aodh Ó Domhnail, Uilliam Burc and Cormac Ó Headghra–Ó hUiginn’s poems often transcend bardic convention and their political, cultural or contractual function as praise or court poems to specific Gaelic families or lords. With a range of expression, subject, imagination, political relationship and music unique among most of his contemporaries, many of his poems mark historical moments and feeling in Gaelic court life, as well as the Gaelic campaign and resistance to colonization by means of singular expressions of relationship with the supernatural, the natural, the mythic, the erotic, the violent and the queer/homosocial. The period of Ó hUiginn’s life which constituted a flourishing in the bardic literary tradition–and his poems, geographical range and relationships throughout Northern Connaught and Southwest Ulster—are the subject of this google digital humanities deep map as an experiment in the radically spatiotemporal, the political and the linguistic plurality of place. As an ecos in flux marked by a sense of chosen relationship or ‘home’ for Ó hUiginn (mo bhaile), this online deep map charts the movement of the poet and his texts in a dynamic rhizome or fractal of early modern Irish land, language, the non-human and inanimate, built and natural environments, history, folklore and toponym—busily functioning despite the upheaval of the Tudor conquest. As such, the mapping project seeks to decolonize early modern Irish spaces and places by freeing an Irish language archive of voices, words and narratives, privileging dynamic Irish relationship with and beyond the human, and demonstrating the power of spatiotemporal unity on the ground in an open digital form. This presentation will offer an introduction to these ideas in English and briefly in Irish with both introductions projected as they exist inside the deep map. Then, we’’ll take a short spatiotemporal journey to early modern Ireland—not as a dead past, but a living, vibrant space—wandering across the island and stopping at different texts as they exist and touch inside the map — birds, their habits and songs, castles and their histories and people, bardic poems and their patrons and poets, excerpts of folklore and historical story taken from the Irish Folkore Commission (duchas.ie), with accompanying images and the place name or toponymy of named Irish place. Our own ‘wandering’ remaps new spaces in relationship with those already mapped online, offering decolonial perspectives on relationships between land, nature and the human. LINK TO MAP IS BELOW

https://www.google.com/maps/d/edit?mid=17o_nT1pkdmGu5P2ABUKhtV0TG4Jht3I

Melissa Buckheit is a graduate student in Irish Studies and English and a poet, translator and dancer/choreographer. Her research interests include early and late modern Ireland, postcolonialism, bardic poetry, the Irish language, Indigenous studies, Place/Space studies, Environmental Humanities, Queer studies and Phenomenology. She recently presented a paper on colonial metanarratives of loss placed in the Irish landscape for ACIS 2024: Ireland at Risk, at Baruch College. She is the author of Noctilucent, among other books, and she has an M.F.A in Poetry from Naropa University. Find her at melissabuckheit.com

What’s Harder for a Computer? Analyzing Meaning vs. Authorship in Al-Ghazali’s Corpus

Ayub Nur (Tufts University)

This talk explores the challenges of applying computational methods to Arabic texts in the digital humanities, using two ongoing projects on Al-Ghazali’s corpus as case studies. The first project examines the term saʿādah (happiness) to determine whether it is used in a technical or generic sense—a seemingly simple question that proves surprisingly difficult to answer computationally due to the small dataset and the term’s contextual nuances. The second project uses stylometric analysis to determine the authorship of disputed texts, a task that is computationally more straightforward because it relies on larger-scale patterns. By comparing these two projects, this talk highlights how the size and nature of the dataset shape what’s computationally feasible, offering insights into the limitations and possibilities of digital humanities research.

Ayub Nur is a senior at Tufts University, completing his B.S in Computer Science with a minor in Digital Humanities. Whenever he is not learning a new language or technology, then Ayub can probably be found making a cup of tea and laughing with his friends. Ayub is currently working on two projects: one at Tufts on text mining the corpus of al-Ghazali, and another at the University of Groningen on analyzing the hadith corpus using AI.

Reconstructions in the Digital Age: Seeing the Past through Virtual Reality

Sean Townsend, Ella Howard (Wentworth Institute of Technology)

In recent years, the popularity of the Meta Quest system has brought virtual reality into over 20 million homes around the world. This widespread and highly accessible technology creates opportunities for historians to share their knowledge of the past with an interested public.

This poster showcases the work being done in the Wentworth School of Sciences and Humanities in the class, Computer Science and Society Studio: Historical Recreation in Virtual Reality. It centers on the experience of the Historical Virtual Reality CO-OP.

In the poster, we detail the class’s effort to recreate the West End neighborhood of Boston in virtual reality, consider the value of effective virtualization of historical data in the modern age, what makes a good reconstruction, and best practice recommendations for both virtualization and virtual reconstruction. Specifically, this poster will detail their work process, the technology used, problems they encountered throughout their development, the sources they used to create the reconstruction, and the effectiveness of their product.

The inaugural Historical Virtual Reality CO-OP employee will share the significant accomplishments he has made throughout the CO-OP, including candid discussion of the iterative nature of these tasks. He will present best practice recommendations for virtual historical reconstructions found through research, as well as the benefits of working closely with knowledgeable people in the field of a reconstruction.

Sean Townsend is an undergraduate student at Wentworth Institute of Technology majoring in Computer Science and Society. He is interested in pursuing a career in museum technology. This semester he is serving as the Historical Recreation in Virtual Reality Co-Op.

Ella Howard, Ph.D., is Professor of History at Wentworth Institute of Technology. She teaches Computer Science and Society Studio, GIS, Boston History, and Environmental History and Impact Assessment.

Framing Marriage and Fertility in China: A Data Science Analysis of Government Discourse in People’s Daily newspaper

Dening Li, Arie Shaus (Mount Holyoke College)

This project examines how the Chinese government frames marriage and fertility issues in the public sphere through an analysis of People’s Daily, the official newspaper of the Communist Party of China. While public discourse on fertility policies has been widely studied, the government's strategic communication of these issues remains underexplored. This research addresses this gap by analyzing 18,782 articles (selected using 10 keywords such as “marriage,” “children,” and “family”) from 2013 to 2024—a period marked by major policy shifts from the one-child to two-child and then three-child policies. Using computational methods—including TF-IDF for term frequency analysis, LDA for topic modeling, concordance in NLTK for context analysis, and semantic analysis with Spacy—this study identifies key themes and shifts in government discourse. Preliminary findings reveal a strategic narrative that portrays marriage as a stabilizing institution and fertility as essential to national progress while downplaying systemic challenges such as gender inequality and economic barriers. By uncovering how the government constructs and communicates these issues, this project contributes to the understanding of the intersection of policy, culture, and public discourse in China’s demographic transition. It also highlights the government’s role in shaping public perceptions and provides a methodological framework for analyzing large-scale textual data in the digital humanities. Additionally, the findings offer insights into the tensions between traditional values and modern demographic trends, particularly regarding women’s autonomy and societal expectations in contemporary China.

Dening Li is a Sociology and Data Science student at Mount Holyoke College, specializing in computational text analysis and public sphere research. Her work explores government narratives on fertility in China, gender identity in women’s colleges, and media representations of marginalized groups. She applies NLP techniques such as TF-IDF and topic modeling to analyze large-scale textual data. Beyond research, she has experience as a researcher, teaching assistant, and community dialogue facilitator. Passionate about the intersection of technology and social issues, she seeks to uncover hidden patterns in discourse that shape public perceptions and policy.

Arie Shaus is a Visiting Assistant Professor in Data Science at Mount Holyoke College and a Research Associate in the Department of Human Evolutionary Biology at Harvard University. His interests include image processing, machine learning, and archaeology.

Heterogeneous stylometric features for premodern languages: a case study in the genres and speaking styles of Latin literature

Joseph P. Dexter (Northeastern University and University of Macau)

Premodern languages present a particular set of challenges for work in the digital humanities, notably sparsity of data essential to the use of modern NLP (Blasi et al. 2022, Sommerschield et al. 2023). Research often depends on creative combinations of traditional humanistic knowledge with computational methods (Dexter et al. 2017, Kestemont et al. 2022). As a model of how to pursue such research on a premodern literary corpus, this paper presents a quantitative approach to analyzing genre and speech in Latin literature. We build stylometric profiles of the canonical genres of Latin prose and verse based on the frequencies of function words, syntactic features, and other non-lexical markers, and use machine learning to identify the characteristic features of each genre (Chaudhuri et al. 2019). After this initial analysis, we then apply the same feature set and dimensionality reduction techniques to analyze progressively finer distinctions: authors writing within the same genre, differences between subgenres, and finally the styles of speakers in individual works. In particular, we present an extended case study of the differences between mortal and divine speech in classical Latin epic. We identify key stylistic features that distinguish human and divine speech and discuss the literary significance of an outlying character appearing in a prominent position at the end of Vergil’s Aeneid. Such use of stylometry differs from customary applications in the domains of attribution and dating, and provides an example of how computation can leverage elements of traditional philology to contribute to literary criticism of a premodern corpus.

Joseph Dexter is a Research Assistant Professor at the Roux Institute of Northeastern University and an Assistant Professor at the University of Macau. He is the co-founder and co-director of the Quantitative Criticism Lab, a cross-institutional digital humanities research group with particular interests in the computational analysis of Greek and Latin literature and in studies of cultural evolution. He received a Ph.D. in Systems Biology from Harvard University and an A.B. in Chemistry from Princeton University.

Enhancing Museum Collection Tagging with Convolutional Neural Networks

Eonbi Choi, Arie Shaus (Mount Holyoke College)

Museums face challenges in efficiently tagging and categorizing their collections, especially as the demand for online access grows while expert resources remain limited. Tags play a crucial role in organizing artworks based on visual elements, themes, and symbolism, helping users navigate collections without requiring specialized knowledge. This study explores the application of convolutional neural network (CNN) models for classifying art images from the Five Colleges Collections Database, a shared resource encompassing objects from six museums in Western Massachusetts. The project evaluates and compares the performance of two CNN architectures, each designed to enhance the classification process. The promising results demonstrate how tagging efficiency can be improved, expanding user accessibility to online collections.

Eonbi Choi is a senior at Mount Holyoke College, majoring in Statistics and minoring in Computer Science. Her interest in computational approaches to art and culture began last year in a Data Capstone class. A visit to the Mount Holyoke Art Museum during the course led her to reflect on how digital tools can increase accessibility and preserve cultural heritage. In addition to her project in digital humanities, she is also passionate about sports analytics.

Arie Shaus is a Visiting Assistant Professor in Data Science at Mount Holyoke College and a Research Associate in the Department of Human Evolutionary Biology at Harvard University. His interests include image processing, machine learning, and archaeology.

From Dugongs to Quetzals to Sequoias: How Global Mythology Quantitatively Embeds Humanity’s Relationship with Nature

Autumn Sehy (Brandeis University)

This poster presents a computational analysis of nearly 7,000 myths and 200,000 extracted flora and fauna from 1000 species, revealing the ecological structures embedded in global storytelling. Combining NLP techniques—including sentiment analysis, affordance and biological-based feature engineering, part-of-speech tagging, and LLM classification—with biodiversity metrics, this research uncovers how mythologies encode utilitarian and functional relationships between humans and nature.

Unlike prior AI-driven folklore studies that focus on structuralist tropes (e.g., the Hero’s Journey) or Eurocentric datasets (e.g., Grimm’s Fairy Tales), this study asks: Which species appear, where, and why? It quantifies biodiversity trends, predator-prey balances, and cultural storytelling constraints across continents, exposing deep-rooted patterns in how humanity encodes nature through myth.

This research began with an accidental discovery: myths can be classified to cultures of origin with Naïve Bayes, logistic regression, and bag-of-words with 87% accuracy—yet when using only mammal features, accuracy plummets to 18%. What drives this gap? By analyzing flora/fauna as cultural signals, this study uncovers striking trends—such as the overrepresentation of predator-prey dynamics and the universal privileging of species with direct human utility (e.g., livestock, medicinal plants, pest controllers).

Created by a former Glacier National Park guide turned computational linguist, this project provides new evidence that mythology is not just storytelling—it is an encoded ecological record, preserving knowledge about the environment and humanity’s relationship with it across generations.

Autumn Sehy is a computational linguistics MS student and former Glacier National Park guide exploring how AI uncovers hidden patterns in folklore. With a background in machine learning, field ecology, and degrees in English literature and anthropology, she merges NLP with biodiversity metrics to analyze how myths encode humanity’s relationship with nature. Her capstone research spans nearly 7,000 myths and 1,000 species to reveal storytelling’s deep ecological structures. When not obsessing over tree communication, she also investigates rhetoric—collaborating with political scientists to analyze the 118th Congress beyond the left-right spectrum.

Intertext.AI: An Augmented Close Reading Interface for Classical Latin using AI for Intertextual Exploration

Ashley Gong (Harvard University)

In the Latin literary tradition, intertextuality—connections between texts that shape meaning and interpretation—is essential for understanding stylistic patterns, thematic nuances, and the relationship between literature, politics, and human culture. However, conducting this advanced humanistic inquiry necessitates specialized knowledge of the Latin canon. To facilitate the development of this domain knowledge for readers, we developed Intertext.AI, a novel digital tool for intertextual analysis that leverages AI to computationally process and search an extensive text dataset. Combining close and distant reading, this system can draw broader literary connections that include voices and texts on the margins of the corpus that receive less attention in existing scholarship, e.g., the comparison of how multiple authors treat women or indigenous populations from antiquity.

Using vector-based semantic search and interactive text displays, users can query Intertext.AI for contextually similar excerpts across the Latin corpus. A pilot study demonstrated that the interface uncovered textual relationships beyond mere lexical matches by visualizing parallels within their broader context. Further, participants ranging from intermediate to advanced classical Latin backgrounds reported greater confidence and ease of use when using Intertext.AI to find allusions than existing digital tools.

Intertext.AI enables classicists to explore intertextuality more efficiently and educators to teach students how to engage critically with a network of texts. By approaching literary analysis with human-AI collaboration, this digital platform encourages wider participation in advanced classical studies for online communities of users and enriches close reading with the computational investigation of unexplored and potentially underrepresented textual connections.

Ashley Gong is a senior at Harvard University studying Classics and Computer Science. Her research explores the intersection of AI and the humanities, focusing on human-centered NLP applications for Latin philology. For her senior thesis, advised by Elena Glassman and Mark Schiefsky, she developed Intertext.AI, an AI-powered interface for intertextual analysis in Latin literature. She is also a research assistant for digital humanities projects in the Harvard Classics Department and Center for Hellenic Studies. Beyond academia, Ashley has interned as a software engineer at LinkedIn and Deloitte, and she will join Watershed Climate as a software engineer after graduation.

Activating Digital Archives: Browsing the Bookshops in St. Paul’s Churchyard

Mary Erica Zimmer, Ian Frankel, Tristan Hoang, Grace Zhang, Bridget Allan (MIT)

In 1990, Peter W. M. Blayney’s groundbreaking The Bookshops of Paul’s Cross Churchyard pioneered the field of “book-trade topography,” visually rendering a vital bookselling district of early modern London by tracing aspects of its mapped footprint, as detailed through intersecting print and manuscript documents. While Blayney’s print-based work revolutionized terms in which research might take place, the affordances of digital media allow us to enrich and deepen the archival interface presented by this complex, layered visualization. By providing access to the document-based foundations of Blayney’s maps, Browsing the Bookshops in St. Paul’s Churchyard excavates and complicates the existing scholarly terrain while developing a dynamic, accessible environment to which further researchers can contribute.

Since 2024, MIT-based undergraduate research teams (UROPs) have engaged the Bookshops project. The range of their research demonstrates this digital environment’s rich potential. Shared by all teams has been the goal of providing access to key documents in multiple genres; here, our transcription and encoding takes place in dialogue with The Map of Early Modern London (MoEML). From this work emerges an interoperable corpus capable of being leveraged to further ends, including studies of gaming, maps and architecture, and the varied, nuanced careers of early modern English female stationers. Our poster will highlight selected developments in the project’s history while bringing forward possibilities for its future.

Ian Frankel is a first-year student within the MIT Concourse Program whose interests include game studies, political science, history, and mechanical engineering. He is in his second semester of work with the Bookshops project and is currently developing an educational board game based on his research.

Tristan Hoang is a first-year student within the MIT Concourse Program whose interests include architecture, maps and mapping, photography, and history. He is in his first semester of work with the Bookshops project and is currently exploring engravings within the early modern English book trade.

Grace Zhang is a first-year student within the MIT Concourse Program whose interests include computer science, mathematics, creative writing, and cultural history. She is in her first semester of work with the Bookshops project and is exploring the work of selected early modern English female stationers.

Mary Erica Zimmer is a Lecturer in Humanities within the MIT Concourse Program. Her work to develop Browsing the Bookshops in St. Paul’s Churchyard has been supported by the National Endowment for the Humanities, MIT History, MIT Literature, the Folger Shakespeare Library, the Bodleian Library, the Digital Humanities Summer Institute, and the School of Advanced Study within the Institute of English Studies at the University of London.

Coding Change: Women and Minorities in Tech

Silvia Rodrigues Follador (Visiting Scholar at WGS MIT; PhD Candidate at FGV EAESP)

{reprograma}, a feminist NGO, tackles Brazil's stark tech industry disparities by providing free coding bootcamps to women, particularly Black and trans women. It is an example of a market-based solution which seeks to foster social justice, decolonize digital practices, and promote social impact in areas overlooked by conventional public policy and governmental agendas. Despite Brazil's growing ICT sector, where 73,000 jobs were created in 2022 (Brasscom, 2023), women, representing 51% of the population, hold only 39% of these roles. Black women comprise a mere 13% of the IT market (Brasscom, 2022), and trans individuals face extreme marginalization, with only 0.02% reaching college (Benevides, 2022). {reprograma} has successfully graduated 1,622 women since 2016, with an 81.1% completion rate and a 63% employment rate within six months. This research investigates how an intersectional approach influences the capacity building of {reprograma}’s participants and in what way(s) does its collective nature affect the participants’ experience. By analyzing this digital education initiative, this research aims to provide insights into how an intersectional framework can foster inclusive and equitable tech environments and how these success stories can be shared with a wider audience.

Currently a Visiting Scholar at the Women's and Gender Studies Program at MIT and a PhD Candidate in Public Administration and Government from FGV EAESP. She holds a Master's Degree from SciencesPo Paris in International Affairs and a double B.A. degree from Pontifícia Universidade Católica de São Paulo and SciencesPo Paris in International Affairs.