| CHALLENGES
The primary challenge in this analysis was dealing with the average dwelling value data. This data is integral to the analysis as it serves to identify the population of seniors that will be able to afford the proposed development as well as eliminate enumeration areas that are primarily non-residential (see Spatial Analysis for further discussion). Unfortunately, the quality of this data is questionable (as discussed in Data Collection) and may have produced unsatisfactory results. "Ground truthing" is required to ascertain the reliability of the results. This image shows average dwelling values for enumeration areas within Vancouver. The dark areas across the middle of the image indicate enumeration areas for which their were no data values assigned (not just really low values) which is problematic as I know there to be residential uses along that corridor.
Initially, I had RECLASSed the average dwelling value data to show stepped values between 0-255 that indicated a level of suitability. The MCE results were not satisfactory however because I had also weighted this data very high. In turn, enumeration areas that had dwelling values that were too low to be considered suitable were identified as suitable. Some of these locations were in known industrial areas and clearly not suitable for a high-end seniors retirement home. I decided that a boolean constraint image, while less flexible, would lead to overall more satisfactory results. The following image shows the results of RECLASSifying the above image.
Another problem I had with the analysis was with the "home" criterion that indicates the location of existing, and potentially competing, seniors homes. When making the data table to import into ArcView for geocoding, I made created an attribute that noted how many units each home had with the expectation of identifying a radius from each home of the area it serviced. This radius would be based on an accepted maximum number of units per senior. However, I found this type of analysis to be impractical to perform in IDRISI . It would require variable buffers based on an ideal maximum seniors population contained within those buffers. Each cell would have to be given a population number based on the density of the enumeration area and assuming that density to be homegenous. Then, buffers would need to be calculated to extend consistently from each home until the population total had been reached. I was not able to find a way to do this in IDRISI. |
||||||||
|
