Katherine McDonough
Katherine McDonough is a Lecturer in Digital Humanities in the Department of History at Lancaster University and Senior Research Fellow at The Alan Turing Institute. Her first book manuscript Public Work: Making Roads and Citizens in Eighteenth-Century France examines unsuccessful socio-economic reforms related to infrastructure development in pre-Revolutionary France. Her digital work uses computational image, text, and spatial analysis to create new ways of doing historical research with large cultural heritage collections as data. With colleagues from Living with Machines, a landmark UK project using maps, newspaper, and census data at scale to write new histories of industrializing Britain, she co-developed the MapReader software library, a tool for analyzing visual information on large corpora of historical maps. She was the UK PI on Machines Reading Maps, a project that creates, curates, and analyzes datasets of text on maps for research and discovery applications is now a Co-Investigator on the Data/Culture project at the Turing, which focuses on building sustainable communities around computational methods and data in the humanities.
Session
MapReader is a software library that was designed for humanities research with big digitised map collections. The winner of the 2023 Roy Rosenzweig Prize for Innovation in Digital History from the American Historical Association, MapReader was developed first within the Living with Machines project, but was created with a wider community of historians in mind as future users. Learn more about MapReader at https://github.com/maps-as-data/MapReader.
MapReader performs two tasks.
1. Patch classification allows users to identify concepts of visual interest on maps, and then to define queries for predicting whether those concepts are present on hundreds or thousands of individual sheets. The power of this approach is its flexibility for any number of spatially-driven research questions.
2. Text spotting makes it possible to create a structured dataset of all text on a map image. MapReader implements models made available during the 2024 MapText competition (Chazalon, Joseph. “ICDAR 2024 Competition on Historical Map Text Detection, Recognition, and Linking”. Presented at the International Conference on Document Analysis and Recognition (ICDAR), Athens, Greece, September 4, 2024. https://doi.org/10.5281/zenodo.13628614.)
This workshop aims to bring together historians and others with an interest in using digitised historical map collections as primary sources for digitally-inflected research. By bringing together peers working in this space, we aim to learn about and discuss ways to encourage open research in the humanities through skill development and shared digital resources and infrastructure.
During the workshop, participants will:
- Learn about the research and theoretical motivations behind MapReader, and how it fits in a growing ecosystem of computer vision tools for humanities research
- Test a demo of MapReader with sample data
- Learn the basics of computer vision and machine learning as applied to computational maps research
- Discuss how to apply MapReader to your own map collections
- Reflect on the opportunities for using “automatic” methods for analysing maps in humanistic research