Database Reference
In-Depth Information
the data and will include understanding the content and the level of detail (scale)
that the data represents: If the spider-sighting dataset only contains sightings to the
nearest 1 km, is this sufficiently precise? Understanding the nature of the content is
a semantic exercise. For example, if the data is of spider sightings, then is the spe-
cies of interest included, and can it be discriminated from the other spider species?
Similarly, if the data describes land cover, then is the right land cover included, and
if it is, is the classification system that is used understood, that is, is unimproved
grass as represented in the data the same as the analyst's expectation and require-
ment? This last point can be quite subtle. What it is really asking is whether there is
(semantic) agreement between what the data collector and the analyst understand to
be unimproved grass. Often, such terms are used in similar but not identical ways,
and if the two definitions are not identical, the analyst then has to ask whether this
will have a significant impact on the analysis. To answer these questions, the analyst
will be reliant on the quality of the documentation and may even have to directly
contact those responsible for the production of the data. There are of course other
considerations when obtaining the data, including such things as permissions and
whether the data is freely available or commercial in nature. But, for our discussion
the important aspects are the locational and semantic appropriateness of the data.
4.3.1.2 Load the Data
Having obtained the data, the next step is to load the data into the GIS for analysis.
If the analyst is lucky, the data will already be in a format that can be loaded directly
into the GIS. When dealing with inherently spatial datasets such as the ones that are
likely to be required for this exercise, there is a reasonable likelihood that this will
be the case, or the data will be in a format that the GIS is able to accept through a
supported translation process. But, it is not always so. Let us suppose for now that
although most of the data can be loaded in a straightforward manner, the invasive
species data has been created using an unfamiliar data format. Now, it is necessary
for analysts to understand this format and to convert the data to a form that is accept-
able to their GIS. Depending on how well the data is documented, this can be a fairly
challenging, though tedious, task.
4.3.1.3 Conduct the Spatial Analysis
The final stage requires the analyst to process the data to test the hypothesis. This can
be achieved by executing the following spatial query: “Locate all invasive plant sites
that are on riverbanks adjacent to unimproved grass and near sightings of the spider.”
This, of course, requires the query to be a little more specific:
Adjacent
may be
interpreted as “within 20 m” to allow for areas where the unimproved grassland may
be close to but not actually physically next to the riverbank. And,
near
may be inter-
preted as within 50 m to allow for the likely travel distance of the spider. The query
is then resolved by the GIS, creating buffers of 20 and 50 m around the river and the
spider-sighting locations and then selecting those invasive species locations that are
contained within overlaps between the two types of buffered areas. Similar queries
can then be executed to find plant locations that do not fit the original criteria, and
by comparing the different sets of results, a conclusion can be reached that either
supports or disproves the hypothesis.
