To prepare the factor data, the first step is to standardize each layer to the same scale, the scale of a map matters to the study because the data must be on a small enough scale to determine population density of census subdivisions, while also large enough to be useable in an MCE spanning all of the Canadian provinces. Because this problem does not require fine detail, it is safe to resample the raster layers to match an appropriate resolution of two square kilometre tiles. The transmission line and cellular tower vector layers will have buffers applied to them to represent effective broadband ranges. Once the factors and constraints are laid out, the MCE will be run which will highlight which populated areas have poor broadband quality. This output will then be clipped with the census subdivisions, which will provide a generalized location that can use improvements.
The optimal connection quality for a populated region is approximately 5.0mbps download and 1.0mbps upload (CRTC, 2018). Fibre Networks will meet this speed easily, while both copper and cellular can vary based on distance and age of infrastructure (Davies, R., European Parliament, & Directorate-General for Parliamentary Research Services, 2015). As such, emphasis will be placed on regions towards the end of the maximum range of the cellular and copper medium by using scaling buffers around each.
The factors are standardized on a scale ranging from 0 - 2048 although the highest values peaked at 1884.61. A scale of 2048 was chosen to create a valuable visualization of the identified regions. This large scale provides an ample variance in colour values and more detail can be shown in the final visualiziation.
The cell size chosen is square metres. This size was chosen because it is sufficient for providing enough detail on a Canada-wide scale without being too computationally expensive. In addition, having a lot of small cells covering a large area provides a more detailed visualization.