Visual GISting: bringing together corpus linguistics and Geographical Information Systems

Corpus linguistics and Geographical Information Systems (GIS) are approaches exploiting computer-based methodologies in the study of, respectively, language and language usage, and spatial patterns in geographical databases. We present an approach that uses corpus methods to bridge the gap between the textual content of a corpus (and, thus, the typically textual concerns of many branches of the humanities) and the geo-referenced database at the heart of a GIS. Using part-of-speech tagging to extract instances of proper nouns from a corpus, and a gazetteer to limit these instances to those representing place–names, a database of the places mentioned in a corpus can be created, visualized, and analysed using GIS technology. It is also possible to visualize the meanings associated with particular place–names, by building GIS databases on the collocation of place–names with particular semantic categories in their immediate context. In this way, we can create maps that visualize the geographical distribution of mentions of concepts such as war, government, or money in a particular data set. The approach cannot be entirely automated and some manual intervention is required. Nevertheless, the method is clearly valuable for the interpretation of spatial phenomena in text corpora.