Database Reference
In-Depth Information
The Semantic Web will not completely solve all problems related to data exchange
and recombination, but it does offer ways to reduce and better manage them. The
Semantic Web also cannot help in all circumstances; it is better suited to certain types
of GI-related problems than others. Equally, the specialist knowledge developed by the
individual communities must not be lost but preserved in the context of the Semantic
Web. The next section identifies where the Semantic Web can help and where it cannot.
4.3 APPLYING THE SEMANTIC WEB TO GI
In general terms, GI has two major uses. GI is an essential element in data analyses
for which location and space are important elements (spatial analysis), and it is also
used to combine data through location. In almost every case, geographic data analy-
sis cannot be performed without the ability to combine data through location, and
elements of the former may also be required to assert that two entities are collocated.
The situation therefore appears to be circular, but can be resolved.
The need to perform spatial analysis was central to the emergence of a GI com-
munity and the creation of GIS. As has already been discussed, GIS mostly operates
through arithmetic calculation performed on explicit geometry, not an area where the
Semantic Web is strong. The emergence of new standards such as GeoSPARQL that
provide some simple geometrical operations for data held within the Semantic Web
do enable simple spatial analysis to be performed based on geometry, but this is still a
long way from the sophistication provided by a GIS. Therefore, the Semantic Web is
limited to simple spatial analysis plus the ability to perform some topological analysis.
On the face of it, this is pretty damning regarding the usefulness of the Semantic
Web to the GI Community or anyone wishing to perform some spatial analysis.
4.3.1 e
xample
Let us suppose the following fictitious situation: Along the banks of a river, it is
noticed that an invading plant occurs unevenly, and the spread is also uneven. It is
postulated that the plant thrives where the surrounding countryside has a cover of
unimproved grass and where there have been sightings of a now-rare spider. Away
from the riverbanks, there appears to be no correlation between spider and plant. It is
decided to perform a spatial analysis to confirm or dismiss this theory.
The actions that are necessary to complete this task are to obtain the appropriate
datasets, load these into GIS, perform the necessary calculations, and present the
results. Expanding each of these tasks, we can gain a better understanding of the
nature of individual processes.
4.3.1.1 Obtain Appropriate Datasets
If we assume that the analyst does not have immediate access to the data, then a
search for the relevant datasets is first required. At present, although a few meta-
data sites exist, the majority of data is not discoverable through directories, and the
process will require a degree of investigation. Having found a prospective dataset
through cunning sleuth work, our analyst must confirm that the dataset does indeed
include suitable data. This process will examine a number of different aspects of
