Spatial dimensions of information have high relevance for everyday problems. A typical example is knowing the locations of the closest stores which have a specific product in stock and are currently open. This geographic dimension information is normally available, but dispersed among a multiplicity of information sources such as isolated Geographic Information Systems, enterprise warehouses, proprietary data formats such as Excel sheets or simple web pages.
The aim of the GeoKnow project is to make information seeking easier by allowing exploration, editing and interlinking of heterogeneous information sources with a spatial dimension. Complex scenarios such as the logistical status of a product within a supply chain and data warehouses of e-commerce systems are also dealt with in the Geoknow project.
Geoknow aims to contribute to the following areas concerned with geospatial data:
- Creation and maintenance of qualitative geospatial information from existing unstructured data such as OpenStreetMap, Geonames and Wikipedia. Geoknow will develop quality assessment methods which anticipate a geospatial search and the acquisition and aggregation of information resources.
- Reuse and exploitation of unforeseen discoveries found in geospatial data. Geoknow will provide methods to acquire, analyse and categorise data that is rapidly evolving, immense, incomplete and potentially conflicting. This will be achieved with:
- Tools and methodologies for mapping and exposing existing structured geospatial information on the web of data, considering comprehensive and qualitative ontologies and efficient spatial indexing.
- Automatic fusing and aggregation of geospatial data by developing algorithms and services based on machine learning, pattern recognition and heuristics.
- Tools for exploring, searching, authoring and curating the Spatial Data Web by using Web 2.0 and machine learning techniques based on scalable spatial knowledge stores.
All these contributions are integrated in the open source GeoKnow Generator framework developed by the consortium. This framework allows the creation of spatial semantic web applications rapidly by:
- integrating geospatial reasoning tools,
- processing of billions of geospatial information sets
- spatial-semantic browsing
- enabling the combination of datasets from the LOD cloud and private data with a geographic dimension.
The GeoKnow Generator will provide a comprehensive toolset of easy to use applications covering a range of possible usage scenarios (e.g. mobility/traffic, energy/water, culture, etc). It will be used by Unister as a spatial-semantic travel e-commerce data management tool; and by BROX as a spatial-semantic collaboration and data integration tool along value-chains in supplier and customer networks. The relevance of these scenarios will be described in another post.