Database Reference
In-Depth Information
Linked Data Web known as the Linked Data Cloud (
Figure 2.1
), where GI-based
datasets like GeoNames
7
and LinkedGeoData (Hellman, Auer, and Lehmann, 2009)
are often more tightly interlinked and act as intermediaries between other datasets.
8
The technical details of Linked Data are discussed further in Chapters 7 and 8,
including a step-by-step guide to publishing geographically based Linked Data. For
now, it is sufficient to say that Linked Data concentrates much more on simple struc-
tures, data discovery, and reuse compared to the focus on expressing and reasoning
over knowledge that characterizes other ontology-based Semantic Web initiatives.
Both aspects of the Semantic Web, however, have significant benefits for the use and
management of GI.
2.4 SEMANTIC WEB BENEFITS
There are several clear benefits of applying Semantic Web technologies to GI:
1. Data integration
2. Data repurposing
3. Data collection, classification, and quality control
4. Data publishing and discovery
2.4.1 D
ata
I
ntegratIon
Currently, it is expensive to combine complex spatial datasets from different data-
bases and difficult to accurately conflate categories of items. The problem is more
than a syntactic one and cannot merely be solved by adherence to one standard file
format. Rather, it is also a problem of understanding what the data means: Is one
organization's definition of a field (the area of land) the same as a different organiza-
tion's? If they are conflated as equivalent entities, what impact would that have on the
results of a database query? An organization interested in the selling and purchasing
of fields will define a field as a parcel of land spatially delineated by barriers such as
fences or ditches, while for a government department interested in providing farm
subsidies, the definition may be subtly different, with the delineation limited to the
extent of crops or accessible pasture. In some cases, the fields referred to may be gen-
uinely the same field in the real world, where for example a field used for pasture has
exactly the same location and boundary in both views. But, differences are also pos-
sible; as an example, consider a field that is not fully planted with crops: The extent
of the field from a land-purchasing point of view will be an area of land defined
by its physical boundaries, whereas the government department could consider the
field to be the area of land covered by crops. Conversely, different worldviews can
classify the same things very differently. A zoologist will define a particular fish in
terms of its species and ultimately relate that back to the tree of life as described by
a taxonomic system starting with domain at the top, moving through kingdom all
the way to species. Seat that same zoologist and the same fish in a restaurant and
worldviews change dramatically; now, the fish is classified within “Seafood” and the
categorization is entirely different. Both classification systems are equally valid but
under very different circumstances. Semantic Web technologies can reduce costs and
