Local search engines are specialized information retrieval systems enabling users to discover amenities and services in their neighborhood (schools, businesses, hospitals, etc.). Local search is a new field of economic activity explored by local business specialists such as Yellow Pages, Yelp or FourSquare, but also and increasingly by major players in the web like Google, Yahoo or Facebook. E-commerce specialists such as Amazon have also declared their interest for local search in the context of service offer, selling for instance services of local plumbers or lawyers. Developing a local search system still raises scientific questions, as well as very specific technical issues. One of the main problem encountered is the partial availability or even the absence of informative contents related to local actors, merchants or service providers. This lack of content makes it hard to design efficient local search systems relying on vector space model. As opposed to traditional search engines that benefit from the indexation of full documents, local search engines mostly rely on business descriptions provided by phone companies or trade registers: names, addresses and telephone numbers are the only available data. Far from adequate to build efficient systems, these basic descriptions must be supplemented by other resources. For instance for a merchant, one would look for more descriptive textual content to help determine its business category (plumber, lawyer...), and extract metadata as hours of operation or specialized field of activity. All these mandatory contents are difficult to collect through traditional information extraction processes. Introducing open data in the architecture of local search engines supports the identification and collection of structured content. Collaborative data such as those made available by the OpenStreetMap Foundation can be of help to identify new dealers, and improve their geolocation. The same data can also be utilized to optimize the mapping systems of local search engines. Semantic Web resources such as DBpedia contain keywords or content for enriching ontologies associated with a local search service. Open data provided by cities or national organizations, such as descriptions of public institutions, opening hours, location of shopping centers are other usable resources. Available open data can be exploited to dramatically improve the design of local search engines and their contents. The purpose of this workshop is to explore new fields of investigation both in terms of algorithmic approaches as well as originality of usable data. The workshop will focus on how open data can be used to enhance the capabilities of local search engines.
Topics of interest include, but are not limited to:Semantic web and open data usage to: