Geosearcher: Location-based ranking of search engine results

Abstract

Many web queries have geospatial dimensions. While online shopping is built on the premise that distance and location are irrelevant (with the possible exception of shipping charges), tourism and onsite inspection of goods have a geospatial dimension and distance and location are relevant factors. Current search engines build indices based on keyword occurrence and frequency for query negotiation using these indices. This approach is fast, robust, and generic but when queries are related to physical locations and distances rather than cyberdistances this approach leaves the user to sort through pages of results. In this paper, we describe an algorithm that assigns location coordinates dynamically to web sites based on the URL. A prototype search system was built using this algorithm that uses this information to re-rank the results of search engines for queries with a geospatial dimension. We found that over 80% of the URLs tested could be assigned correct location coordinates. This work makes a contribution to retrieval on the web by providing an alternative ranking order for search engine results so that users with queries with a geospatial dimension can more readily use the results of general search engines rather than special purpose applications.