Various approaches were undertaken during the development phase of this project to identify best suited representations of population clusters within the GIS. Approaches were based on vector and raster database models and were tested in the same environment to see what advantages and disadvantages each held over the other.

In one example, raster cells were created based on union of the DA centroids with a point grid, which was generated based on point distance from the nearest DA centroid. This technique allowed the cartographer to visualize the GVRD in a series of evenly spaced grid cells (1km²). Other adaptations of this method were employed to gather information concerning distances from population centroids to health care facilities and transportation routes. These methods proved advantageous for traditional multi criteria analysis, thus allowing the cartographer a glimpse of the GVRD under many, and often opposing, situations (click here for a sample raster basemap). In addition, interpolation techniques were undertaken in order to identify high density areas of certain criteria(s). This method proved advantageous to visually recognize areas within the GVRD that involved high density clustering of like-attribute data.

These approaches, although informative, were not without their disadvantages. Some pitfalls of this technique were observed in the allocation of clusters using the GRID approach. In certain events, grid clusters generated through interpolation, although useful to observe high density, would place clusters in uninhabitable regions of the GVRD, such as in rivers and inlets. This stemmed from the interpolation process, which did not take into effect geographical land barriers. Equally important was data compatibility and entity identification. The advantages of vector data to hold multiple attributes and to be layered with other data layers proves much more useful then traditional logarithmic mapping techniques (for the purposes of this project).

Mapping the spatial attributes of the GVRD Census tables through point data representation is a delicate matter. In order to make clear map products centroids were created from the DA polygons. This allowed the cartographer to model the information in an ordinal scale using a standardized gradient technique to represent different DA's according to a pre-determined value (age, density, etc.). The creation of centroids, although beneficial for visual comparison, is inherently error driven as it transforms the location of the entire population in a given area (km²) by averaging the number of X and Y vertices in the polygon and to produce a point that lies somewhere in the mean center of the area. One has to ask if the entire population lives within a single point of a DA? This problem is only intensified in DA's that are of large areas versus those that are located in the downtown region and generally much smaller in size.

The power of maps to manipulate and transpose information according to the views of the cartographer is a powerful skill and one that demands a strong understanding not only in map making, but also in economic, political, and geographic thought. This project, although goal focused with producing a useable product, also took into consideration the underlying techniques that go into the production of a map.

Ultimately, it was found that both data models were composed of various capabilities for obtaining and comparing spatial data in relation to this project. To ensure accuracy, the data validation techniques were repeated under each mapping stage in order to transfer the attribute tables of the the GVRD Census tables into spatial form. Below are two sample maps that were created to determine how centroids were to be represented in the 1:12 000 scale maps.

vector model raster model