Due to limitations in time and data, this study rests on a census resident population measure to determine the demand for bike stations. This was based on some studies pointing to the importance for residential users to be proximal to bike docking stations, which influences use (Bachand-Marleau et al., 2012). However, there are significant other measures that need to be factored into who will use the stations. The most obvious other population relevant to the study area, apart from residents, include the high number of employees present during the day which commute into the downtown from outside. A more comprehensive analysis would consider the daytime population as opposed to just the nighttime population to determine demand.
Tourists will also use the bike-share system, of course, taking into consideration the density of attractions, restaurants, and parks (note: Stanley Park). However, I have not incorporated this into my study, considering the City wants to target residents and commuters to reduce competition with local bike rental businesses (see here).
Because my location-allocation analysis is entirely based on these sorts of demand factors, there tends to be less points stationed around low-demand areas and more stations around high-demand areas. This becomes problematic if the dispersal of stations is intended to be consistent, with an even amount of stations spread throughout the area. The assumption in this study is that this is not the case, and that stations will be strategically located. In this sense, this study represents just a preliminary sketch of where the stations could potentially be located based just on demand in terms of resident population.
With regards to the bikeability analysis, one of the major flaws came with the reclassification of each factor from a continuous raster to a scale of 1-10. In order to conduct a weighted overlay analysis to create a suitable bikeability map for the area, layers had to be reclassified to a common scale; however, this process introduced a high amount of subjectivity on my part in determining which classification scheme to use, which in turn, influenced how significant each layer was.
If you notice any other flaws, don’t hesitate to contact me and let me know. I welcome and greatly appreciate any eedback!
