Theme

There is a vast quantity of information in text documents and other media that is referenced to geographic space. The discipline of Geographical Information Retrieval (GIR) is concerned with developing methods to gain access to this geographical information, with a particular focus on the content of web documents and social media. Because much of the information is in the form of unstructured or semi-structured text, there is a challenge to develop methods that can automatically recognise and interpret the geographical terminology and spatial or spatio-temporal concepts that people use when recording and querying the information. GIR falls at the intersection of Information Retrieval (IR) and Geographical Information Science (GIScience) resulting in research and systems development that benefits from the fusion of text-based methods for information extraction, natural language processing, indexing and search with GIS methods for spatial data management, analysis and visualization.

The workshop invites contributions on the following topics, and other research related to GIR:

      Detection, disambiguation and geocoding of geographical references in text;
      User needs for geographic search;
      Classification of web documents and social media with regard to their geographic foci;
      Interpretation of spatial natural language in documents and queries;
      Extraction of geographically-specific facts and events from text documents and social media;
      Spatial and spatio-temporal indexing of documents and other media objects;
      Modelling, construction and integration of ontologies, gazetteers and geographic thesauri;
      Reasoning with geo-spatial facts for purposes of information retrieval;
      Geographical query interfaces for search on the web;
      Geographic question / answering systems;
      Geographic search engine architectures;
      Relevance ranking of geographical information;
      Evaluation methods for geographic search.

We invite both long papers (8 pages) and short papers (2 pages). Long papers are expected to report on relatively mature research results, while short papers may cover more speculative or early stage research that will stimulate discussion at the workshop.