Geographic Information Retrieval and Visualization of Online Unstructured Documents

Newspapers, travel narratives, blogs, books and the Internet hold a huge amount of geographic information that can be extracted in order to provide visual exploration. Also, the understanding of place references involves knowledge of the document context. In this way, the study of tools for disambiguation is needed. For the automatic annotation of time and location, both shared world knowledge and document context needs to be captured. This paper is centered on analyzing online unstructured documents: travel narratives and online newspapers. Our approach is based on the exploration of tools able to make automatically the disambiguation of placenames. In this case, we have used a Geoparsing Web Service to extract geographic coordinates from the online unstructured documents. Once geographic coordinates are extracted by using eXtensible Markup Language (XML) we draw the geo-positions and link documents into a map image in order to visualize textual information.