The National Endowment for the Humanities and the UK-based Arts and Humanities Research Council recently announced grants in support of Machines Reading Maps: Finding and Understanding Text on Maps, a transatlantic collaboration that brings together expertise from the USC Libraries, the USC Spatial Sciences Institute at USC Dornsife College, the Alan Turing Institute and British Library’s Living with Machines project, the Digital Insight Lab at the AIT Austrian Institute of Technology, the Library of Congress, and the National Library of Scotland.
The collaborative project will develop new machine learning tools for extracting data from large-scale historical map collections and combining it with other geospatial and historical data to create a fuller and richer understanding of place. Machines Reading Maps addresses a principal challenge for libraries, archives, and other cultural heritage institutions with vast map collections—ensuring that these valuable troves of historical information can enter meaningfully into scholarly and public conversations about the complex stories of how Los Angeles neighborhoods or British cities were shaped by social forces like red-lining and segregation in the U.S. or England’s industrial transformation during the 19th century.
Deborah Holmes-Wong of the USC Libraries and Yao-Yi Chiang of the Spatial Sciences Institute at USC Dornsife College will be working with Katherine McDonough and other collaborators from the Alan Turing Institute and the British Library’s Living with Machines project; Rainer Simon of the AIT Austrian Institute of Technology; and partners at the British Library, Library of Congress, and the National Library of Scotland. The work builds on prior collaborations between Holmes-Wong, Chiang, and Zahid Rafique and Wayne Shoaf of the USC Digital Library and Chiang’s pioneering work on the Strabo open-source map processing software and techniques for unlocking historical information from maps.
The team will work with a vast collection of late 19th and early 20th century Sanborn Fire Insurance maps recently digitized by the Library of Congress and troves of Ordnance Survey maps held by the British Library and National Library of Scotland, Goad fire insurance maps held by the British Library, and a number of 17th to 19th century map collections such as the British Library’s King’s Topographical Collection and the National Library of Scotland’s collection of Scottish town plans.