Thematic data extraction from Web for GIS and applications

Many Web pages contain textual descriptions about locations such as addresses, phone numbers, landmarks, and names. These location-related descriptions are valuable geospatial information for business applications. Therefore, the Web can be perceived as a large geospatial database that could provide up-to-date data for GIE. Currently this rich and frequently updated Web geospatial information is underutilized. Most related work has been focused on identifying and extracting location names from Web pages for the purpose of page indexing. Little research has been found on extracting and using various types of Web geospatial information for enterprise GIS applications. This paper tries to fill the gap by (1) proposing new algorithms for retrieving different types of geospatial information on Web pages and resolving location name ambiguity issues; (2) exploring how GIS can leverage extracted Web geospatial data for enterprise applications.