Resources

We’ll maintain here a running list of resources at the intersection of . . .

text here — Image created by Nick Budak. Inspired by the first issue of Computers & the Humanities, a journal that ran from 1966–2004.

. . . particularly those that focus on 1) the uses of machine learning in humanities research, and 2) critical, historical, and ethical frameworks for analyzing the consequences of machine learning techniques being deployed in the world today.

Readings 📖

A bibliography we recommend is the Critical Dataset Studies Reading List, run by the Knowing Machines project.

Data Sets & Data Work 💾

The Center for Digital Humanities publishes datasets through research partnerships and curation grants. They can be viewed here. The CDH also curates a list of Humanities Datasets in Context, organized by discipline and annotated with information about data format and collection practices.

Tony Chu and Stephanie Yee, “A Visual Introduction to Machine Learning,” R2D3.

Amy A. Winecoff and Elizabeth Anne Watkins, “Artificial Concepts of Artificial Intelligence: Institutional Compliance and Resistance in AI Startups,” in Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society, 2022, 788–99.

Evan Odell’s Hansard Speeches 1979-2020. Version 3.0.1

Campus Resources

Princeton Research Data Service, for “expert services and infrastructure to store, manage, retain, and curate digital research data.”
Princeton University Libraries Data and Statistical Services
Princeton University Art Museum API Documentation
Education and Training events run by Research Computing

General Portals for Humanities Data

Matthew Lavin’s humanitiesdata.com contains a running list of datasets and recipes
Datasets list maintained by Melanie Walsh, broken down by subject and including example uses and tutorials for each dataset.
Alan Liu’s DH Toychest, list of demo corpora (text collections ready for use)
Journal of Open Humanities Data, peer-reviewed forum for reports on the curation and publication of new datasets. Articles describe “humanities research objects or techniques with high potential for reuse.”
Rutgers University Libraries’ “list of collections and datasets that are (mostly) distinctive or unique to Rutgers, and which may serve as excellent source material for digital humanities research or pedagogical projects.”
UC Berkeley Library’s Text Mining & Computational Text Analysis web portal, with tutorials, sources, etc.
See also this conversation on how to define and understand “humanities data.”

Tools ⚒️

The following tools and platforms were mentioned throughout the institute:

Notion, a knowledge management platform for rich spreadsheets, where each row can become a document or a form. Also has task and project management features for collaboration.
Regex 101 for learning about and testing regular expressions, or, commands for searching for various types of text.
Tropy, an app for organizing photos, notes, and bibliographic metadata for archival researchers. Developed by the Roy Rosenzweig Center for History and New Media at George Mason University.
Obsidian, a plain text / markdown notetaking app that allows you to hyperlink between notes, or add #tags. Similar to a personal wiki for all of your notes. Has lots of plugins for literature citation notes, metadata, daily task lists, etc.
Altair, a visualization library for Python, based on Vega, developed by the University of Washington Interactive Data Lab. Vega is “a visualization grammar, a declarative language for creating, saving, and sharing interactive visualization designs.”
Voyant, a basic dashboard for text anslysis and an off-the self alternative for exploeratory data analysis.
Distant Viewing Toolkit, a Python library for performing machine learning on still and moving images, developed by the Distant Viewing Lab at the University of Richmond.
HiPSTAS Audio Tagging Toolkit, a Python library for labeling audio data for training machine learning classification models, developed by Steve McLaughlin, in the High Performance Sound Technologies for Access and Scholarship Project at University of Texas, Austin.

Technical Instruction ⚙️

Python Notebooks & Textbooks

Quinn Dombrowski curates a list of Jupyter Notebooks for digital humanities. These notebooks, in the spirit of “literate programming,” contain data files and executable code interwoven with textual descriptions of what’s happening. Dombrowski also curates a list of resources specifically for multilingual NLP.

Peace Ossom-Williamson and Kenton Rambsy, The Data Notebook (Mavs Open Press, 2021), a platform-agnostic “online suite of open interactive resources” on the principles of data storytelling.

Nick Montfort, Exploratory Programming for the Arts and Humanities (Cambridge, MA: MIT Press, 2016).

Stéfan Sinclair and Geoffrey Rockwell, The Art of Literary Text Analysis (2018), a series of linked Jupyter notebooks.

Folgert Karsdorp, Mike Kestemont, and Allen Riddell, Humanities Data Analysis: Case Studies With Python (Princeton University Press 2021), a print book and an online Jupyter Book. Includes an introduction to probability.

Melanie Walsh’s Introduction to Cultural Analytics and Python [book-length course]

W.J.B. Mattingly’s Introduction to Python for Text Analysis [book-length course]

Workshops

Most of these workshops are now conducted online via Zoom and require an application and registration fee. The CDH offers Graduate Training Grants to offset the cost of attending these workshops.

TAPI: Text Analysis Pedagogy Institute (JSTOR Labs / Constellate), June - July
Programming 4 Humanists at Texas A&M, March – May
DHSI workshops, June
Santa Fe Institute: Foundations and Applications of Humanities Analytics, July
Leuven + Leiden DH Summer School, June
University of Oxford DH Summer School, July

Online Tutorials

Be sure to check out Foundations and Applications of Humanities Analytics, cotaught at the Santa Fe Institute by David Kinney, one of the instructors at our Institute. This Foundations course “aims to empower scholars in the humanities by eliminating the black box of computational text analysis.” All lectures are available as a YouTube playlist.

CUNY’s Digital Humanities Resource Initiative (DHRI) includes guided tutorials on the following topics:

Programming Historian features lots of excellent peer reviewed workshops in English, Spanish, French, and Portuguese. Some highlights from the 91 lessons currently available in English include:

Library Carpentry’s lessons are geared towards library and information science professionals. They include:

BERT for Humanists provides tutorials in natural language processing (NLP) and machine learning for a humanities audience. Workshops (with Google Colab Notebooks) include:

Constellate is a platform for text analytics developed by ITHAKA (the organization behind JSTOR). Tutorials include:

Finally, some miscellaneous tutorials we’d recommend include:

Brandon Walsh and Rebecca Draughon A Humanist’s Cookbook for Natural Language Processing in Python
Lincoln Mullen’s Computational Historical Thinking with Applications in R
Dennis Tenen’s Intro to Operating Systems and Command Line Interfaces (CLI)
Matthew Lincoln’s workshops
- Critical Data Visualization with Palladio
- Tidy Humanities Data
Elements of AI, “a free online introduction to artificial intelligence for non-experts.”
If you want to train your own NLP Model for Named Entity Recognition (rather than using the pre-trained model):
- W.J.B. Mattingly’s tutorial on how to create a training dataset for spaCy
- W.J.B. Mattingly’s tutorial on how to train your own NER model in spaCy

Examples of Humanities Research Using AI & ML 🧪

DeepMind’s Ithaca, a project using AI to restore and attribute texts from Ancient Greece using neural networks as an aid to epigraphy.

MIT researchers using AI to decode the still-undeciphered Indus script. The paper published in Science concludes that the script is likely not just symbols, but language.

AEOLIAN (Artificial Intelligence for Cultural Organisations), including this writeup on use of AI with medieval manuscripts.

Ben Lee’s video walkthrough of the Newspaper Navigator ML tool and the data paper behind the Newspaper Navigator dataset, “Compounded Mediation: A Data Archaeology of the Newspaper Navigator Dataset“