Objective 2: To Develop a Relevant GIS-based Model that can be used to Evaluate the State of Canada’s Broadband Network.
An MCE is a multicriteria evaluation which uses raster data comprising several maps, in each map, all cells (pixels) are assigned a mathematical value based on criteria and constraint weights. MCE’s are used because they are the most optimal for combining qualitative and quantitative data to find the best possible locations for improvement (Eastman, 1999). Using an MCE for this process is imperative because the project revolves around the spatial exploration of Canada and aims "to investigate a number of choice possibilities in the light of multiple criteria and conflicting objectives” (Voogd, 1983: 21). Using an MCE is beneficial to the study topic by introducing a useful tool for the allocation of activities and spatial modelling which is important when conducting research into broadband networks while simultaneously weighing the application of a decision, for instance, the affected population, against multiple other criteria.
For the purpose of this analysis, three separate MCE’s will be developed. Each MCE is a different type of broadband infrastructure; Wireless, Copper, and Fiber. Having 3 MCE’s is beneficial for describing how different infrastructures have varying levels of quality in different regions across the country. Each of the MCE’s has similar data sets to represent each factor but there is some differences. The similar layers include; distance from urban centers, the population count/densities, and population ecumene. However, the upgradability and existence layers change depending on the technology.
Each MCE will have also have slightly different weighing for their corresponding factors. This is required because the varying broadband delivery methods have different functional requirements. These weights are determined using the Pairwise comparison method. This method is advantageous for determining weights when a small number of factors are being considered.
The general MCE equation used is as follows, where W is the weight of the corresponding factor:
"Inhabitation" * (("Centres" * Wcentres) + ("Upgrade" * Wupgrade) + ("Exist" * Wexist) + ("Population" * Wpopulation))
The Equations have been taken straight from the Python code and represent the changes in each MCE, while using the same layers, the wireless script is as follows:
arcpy.gp.RasterCalculator_sa('"exists_layer.tif" * ((("eucdist_final.tif" / 50000) * 2048 * 0.3889) + (("population_layer.tif" / 4900) * 2048 * 0.0687) + (("urban_layer.tif" / 100000) * 2048 * 0.1535) + ("Wireless_Fix.tif" * 2048 * 0.3889))
For the wireless MCE, upgradability is represented by the location of cellular towers across the country, and the existence of broadband infrastructure is represented by the CRTC’s fixed wireless and LTE availability datasets. As a result, both these factors hold heavy weights, it should also be noted that the population and density factor has a low weight, this is because of the vast 50-kilometre range that cellular towers can support. (Table 1.)
The following equation is taken directly from the Python script used to automate all of the MCEs, the equation for Copper broadband is as follows:
arcpy.gp.RasterCalculator_sa('"exists_layer.tif" * (("Roads_Euclidean.tif"/4975.66 * 2048 * 0.0789) + ("Copper_Fix.tif" * 2048 * 0.5193) + ("population_layer.tif"/4900 * 2048 * 0.2009)+("urban_layer.tif"/100000 * 2048 * 0.2009))
For the Copper wire MCE, upgradability is represented by a 5-kilometre buffer around each communication line, and the existence of broadband infrastructure is represented by the CRTC’s existing copper line dataset. The 5-kilometre range of copper wire means that it is imperative that new broadband infrastructure is within the pre-existing infrastructure, this is the primary reason for the heavy weight of the Existence of Broadband Network factor. (Table 2.)
The following equation is taken directly from the Python script used to automate all of the MCEs, the equation for Fiber Optic broadband is as follows:
arcpy.gp.RasterCalculator_sa('"exists_layer.tif" * (("Roads_Euclidean_Fiber.tif"/25000*2048*0.0356) + ("Fiber_Fix.tif" * 2048 * 0.3904) + ("urban_layer.tif"/100000*2048*0.3904)+("population_layer.tif"/4900*2048*0.1837))
Finally the Fiber MCE, upgradability is represented by a 25-kilometre buffer around each communication line, and the existence of broadband infrastructure is represented by the CRTC’s existing fibre line dataset. The installation of fibre is contingent on the existence of fibre nearby, as a result, the factor that corresponds has a large weight (Table 3).