The Hermitage Plantation Guest Register project demonstrates how AI-assisted transcription can unlock geospatial data embedded in difficult-to-transcribe historical documents. Using a customized large language model and a field-based transcription platform in combination with traditional humanistic research methods, the authors transformed a 159-page handwritten register of plantation tourists from the 1930s into a structured dataset, enabling analysis of visitor origins and travel patterns. The resulting workflow offered new insights into the history of early plantation tourism and models a replicable, low-barrier method for extracting geospatial data from handwritten archival materials.
Citation
Philip Mills Herrington & Loren Moulds, Reading the Register: Using AI-Assisted Handwritten Text Recognition (HTR) to Unlock the Hermitage Plantation Register of Guests, Journal of Map & Geography Libraries (2026).