Suggestive geo-tagging assistance for geo-collaboration tools

An argumentation map is an online discussion forum for spatially related topics that combines the forum with an interactive map. The utility of an argumentation mapping tool highly depends on the accuracy and quantity of the geo-tags that link the discussion contributions to geographic locations. These geo-tags can be created manually by the users of the argumentation map or automatically by a geo-parsing application. However, in the case of manual geo-tagging users often do not create geo-tags as extensively as desired. In contrast, automatic geo-parsers have difficulties with the informal language often used in user-generated content and with resolving small-scale features. This paper proposes a hybrid approach for geo-tagging user-generated content which involves the users in the process but supports them with an automatic geo-parser which suggests locations. The implementation of a prototype as well as a human participants test are presented in order to analyze the geo-tagging performance of this approach. It turns out that it is possible to reduce the number of geo-tagging errors but keep the recall rate approximately constant, compared to automatic geo-parsers.