Project Bamboo has launched Places-Text, a feature demonstrator to illustrate how the possible service in Bamboo’s infrastructure could give researchers analytic tools to identify places mentioned in texts, including books, journal articles, and Web pages. Because many scholars regard place as a key concept in historical and archaeological studies, services to algorithmically identify places in large corpora can open many new and exciting research opportunities.

The following video provides a brief introduction to a possible Project Bamboo Places-Text service.

This discussion of Places-Text services is, in large part, based on the work of the Google Ancient Places project (GAP), led by Elton Barker, Leif Isaksen, Eric Kansa, and Kate Byrne. GAP is attempting to demonstrate the feasibility of automated approaches to the identification of ancient places mentioned in texts, drawing upon public-domain books from the Google Books corpus. To do this, GAP uses the Pleiades Gazetteer of the ancient Mediterranean world (mainly Classical Greek and Roman places) and the Edinburgh Geoparser to algorithmically identify places in unstructured text. GAP efforts have been key to understanding and shaping requirements for deploying place identification services within the Project Bamboo services infrastructure.

Additional information about the Places-Text service, including a mock-up UI, analytic results, and an example map are available on the Places-Text demonstrator site. Project Bamboo will be adding more demonstrators of existing and possible applications and infrastructure throughout the project.