Project title: Deriving vernacular geography through Machine Learning from social media content, investigating efficacies of the approach and potential uses of this information while protecting personal identity, privacy and freedoms
Departments: Law and Computer Science
Contact: Peter Adey
Project outline: References to places or locations permeate many aspects of day to day living and professional activities, using a mixture of what might be called ‘official’ or formal place names as might be found on a published map or gazetteer, and informal or alternative names.
The latter may be termed ‘vernacular geography’ and are names that, for example, people local to an area may use, or that may be in use by certain social or community groups. Knowledge of these names, and the areas or locations to which they refer, is significant in, for example, responding to emergency calls or improving web search and information retrieval operations using place names/identifiers.
Ordnance Survey, the national mapping agency for GB, is interested in approaches to identify vernacular place names and the approximate areas or locations to which they apply, in order to improve gazetteer data nationally. Also Ordnance Survey is interested in exploring applications of advanced analytics using spatio-temporal social data.