William Ward

Machine Learning for Natural Habitat Change From Historic Maps

Completion: 01 JAN 2027

Background

BSc Geography

PhD

Machine Learning for Natural Habitat Change From Historic Maps

My research focuses on designing scalable methods to create spatial maps of past landscapes from extensive and underutilised collections of historic maps. I am looking at the best ways to leverage computer vision models based on neural networks to automate the conversion of these scanned maps into structured data. This allows information of historical landscapes to be systematically integrated into the same spatiotemporal models used for modern landscape planning, providing new ways of studying the development of extant habitats from data in the humanities. By focusing on underutilised, large-scale surveys of England and Wales from before 1850, I aim to improve methods for geospatially modelling human-environment relationships and ecological change at the localised scale further back in time.

Supervisors

Lead Supervisor
Leif Isaksen

Digital Humanities · Faculty of Humanities, Arts and Social Sciences
Co-Supervisor
Zeyu Fu

Department of Computer Science · Faculty of Environment, Science and Economy

External Partners

Matthew Heard, National Trust
Tom Dommett, National Trust

Back