Exploring Historical Links between Scotland and India using Geoparsing

Abstract

A significant amount of spatial information can be derived from unstructured datasets available in web pages, e-books, and digital archives. When attached to a geographical perspective, this information can bring a whole new dimension to any story. Geoparsing is one such concept that is very useful in extracting spatial data from any unstructured text source. The extracted location is then geo-tagged against a standard spatial corpus, usually a gazetteer. Geoparsing complemented with Natural Language Processing algorithms can effectively automate this process of identifying and geo-tagging the extracted spatial data. The research illustrates geoparsing from a historical perspective by extracting place names from a corpus of biographies of famous Scots who travelled to India from the 18th to the early 20th Century. The study provides an effective workflow for extracting and georesolving place names from historical archives. The results identify spatiotemporally significant regional patterns over a period of 240 years (1707-1947). The findings highlight the potential of merging demographic characteristics with georesolved information. The research concludes that geoparsing combined with automated workflows for data management provides a viable method to study spatiotemporally significant studies. The lack of objectivity associated with historical studies should not discourage the use of machine intelligence for such studies; future studies should instead seek to exploit the full potential of machines to benefit the discipline of Historical GIS.