Porter, Catherine and Gregory, Ian (2015) Geographical text analysis:approaches to understand the geographies in texts. In: Association of American Geographers 2015, 2015-04-212015-04-25, USA. (Unpublished)
Full text not available from this repository.Abstract
Many texts carry a wealth of information about places and geographies, however, GIS and spatial analysis as currently conceived are largely concerned with quantitative and cartographic form and thus ignore textual sources. Much of the currently material becoming available to scholars in digital form is textual, ranging from digital archives to born-digital sources. These corpora frequently contain millions, if not billions, of words of text. There is thus a pressing need for geographers to develop techniques that allows them to do this. This paper explores an approach we term Geographical Text Analysis (GTA) that involves combining approaches from GIS and corpus linguistics, two fields that rarely communicate but have much to offer to each other. GTA involves going through several stages: firstly place-names within the text need to be automatically identified and associated with a coordinate-based location. Doing this allows the corpus to be mapped. From here, more sophisticated analyses can be undertaken that combine GIS-based spatial analysis techniques that allow us to ask 'where is the corpus talking about?', with approaches from corpus linguistics that allow us to ask 'what is the corpus saying about these places?' The paper draws on a range of different examples including a relatively small corpus on the literature of the English Lake District, and Early English Books Online (EEBO) a much larger corpus from which we explore the geographies of prostitution in the seventeenth century and before.