In-Depth Cartographic Representation
and Geographic Information Types
Most cartographic representations work with GI stored on a computer. Gener-
alization describes techniques for changing the cartographic presentation. It is
a type of transformation, but focuses on graphical transformations (although,
if used later as GI, the graphical changes will become part of the GI). The
types of GI greatly inf luence the cartographic representation choices. Ideally,
these issues are addressed when first collecting the GI. The desired forms and
formats of the cartographic output can be considered and data correspond-
In many cases, however, the cartographic representation relies on GI col-
lected for other purposes. Vector GI representation types are most common,
as they best ref lect traditional cartographic practices. Points will be used for
the cartographic representation of things and events at small scale; at large
scales they may be represented as areas—for example, as towns, buildings, or
monuments. Lines of vector and network GI types usually remain lines, but
may be strongly simplified at larger scales. This can be a major limitation in
representing network relationships. Conversely, raster GI representation types
(of both discrete and nondiscrete GI) may need to be aggregated at smaller
scales to highlight key properties.
GI types can always be transformed to other types. This often provides an
adequate solution, but care needs to be taken in order to assure that the accu-
racy of the new type will not suggest conclusions that lack support in the origi-
nal GI. Streets, for example, are often made available as street centerline data.
The actual width of the road is not indicated by these lines, nor is the attribute
for the street width usually included. A generic width can be assumed based on
street type, but clearly buffering the street centerlines by this distance is an esti-
mate at best in most cases and erroneous at worst. A note explaining the pro-
cess of transformation and pointing out the error should be included and dis-
played on any maps made with the transformed GI.
tion is also needed because of emphasis on certain thing, events, or relation-
ships that are central to the communication role of the map or visualization.
Work on generalization in cartography recognizes a number of different
operations. The typologies of generalization operations vary, but certainly
distinguish between points, lines, and areas. Other issues reflect institu-
tional, functional, or conventional understanding of the cartographic repre-
sentation and geographic representation that the generalization should sup-
port. To introduce the generalization operations, the five most
commonplace operations offer sufficient insight into the significant role of
generalization for cartographic representation and communication.
Aggregation is used to reduce complexity. It can either be used to merge fea-
tures that border one another (e.g., land use types) or to group feature that