Page 294 - Contributed Paper Session (CPS) - Volume 4
P. 294
CPS2230 Braden P. et al.
polygons in an existing shapefile carry no meaning when filled with color due
to a small size or the location of the polygon.
While there exist various software solutions for simplifying the boundaries
of shapefiles, this has never been done on a large scale. Further, finding
simplified boundaries for a given country that are ready to use poses a
different challenge in that the simplified shapefiles available for public use are
not available in a single location, but are scattered throughout various
packages in R and publications.
The ultimate goal of this article is to provide access to shapefiles for as
many countries as possible to allow for wider use of spatial maps, particularly
linked micromap plots. In addition to the recommended modifications and
shapefiles corresponding to them, there should be tools accessible to users to
further modify shapefiles in a manner that would be more meaningful for
them.
4. Methods Used in R:
There are several types of modifications that are available for addressing
the problems caused by the underlying geography of a region. Some of these
modifications will require the removal of certain polygons, enlarging areas that
are otherwise difficult to see when compared to larger surrounding areas, or
manipulating the shape altogether by moving a given area to a location that
may be more meaningful.
4.1 Thinning Boundaries
In order to understand why the process of thinning needs to be applied to
a shapefile, one needs to understand how a shapefile is constructed. At first
glance, a shapefile is simply the outline of some region, often a geographic or
political boundary. These outlines are created by connected line segments
tracing around the borders. The lines themselves can be considered as a dense
collection of points.
Shapefiles tend to be more complicated when the administrative
boundaries are created to follow the geographic features in the area. This is
especially true in coastal regions where the boundaries tend to not follow a
straight line. When plotting such regions, this can be an intensive and time-
consuming process for a computer. In order to simplify these boundaries,
thinning is often the first step.
The process of thinning is as intuitive as it sounds. If we consider the many
individual line segments that make up a shapefile’s boundaries to be a dense
collection of points lying next to each other, then the process of thinning is
selecting a proportion of these points to keep in our final shapefile.
Shown in Figure 1, there are two outputs for Mexico using two different
proportions of total points to keep. These shapefiles also act as an example of
283 | I S I W S C 2 0 1 9