These are the images used to arrive at my final analysis:
Constraints:
My landuse constraint file. I considered Open/Undeveloped, Residential-Townhouse, Lowrise/Highrise Appartments, Commercial, Residential/Commercial Mixed, and Industrial landuse types as the only landuses suitable to build on.
My aspect constraint file. I considered all aspects that were between 90 and 270 degrees as the minimum suitability for a pub with a patio.
All below images on the left are before using the Fuzzy module. All below mages on right are post Fuzzy, and thier legend shows a scale of 0 to 255. These maps show the suitability of an area, ranging between lowest suitability (0) and maximum suitability (255):
Factors used for my analysis.
1)South Facing Slopes
The Aspects image of Burnaby. It was fuzzied using a symmetric sigmoidal function. I gave the higest suitability to slopes between 135 and 225 degrees. All aspects 45 degrees on each side were given a diminishing suitability until the aspects started becoming northerly, ater which they were given a suitability of 0.
2)Landuse Factor
The Landuse factor image. I scaled it so that all favourable landuses had values between 1 and 650, and gave all undesireable landuses a value of 2000. Then, with the fuzzy module, I used a J-Shape function to give decreasing suitability as the value of the factor increased.
3)Major Roads Factor
The Major Roads factor image is based on the idea that the further from a major road, the less likely the idea that a pub would be exposed enough to generate revenue. Thus, in my fuzzy image, a sigmoidal function type was used to give a higher suitability closer to a main road, and at 400 meters, suitability was considered to be 0.
4)Skytrain Distance Factor
This image is based on the fact that it is better to take transit than it is to drink and drive. I assumed that most people are willing to walk about 20-30 minutes to get to a location beyond a Skytrain Station, and based on my own experience, I estimated that to be about 1.2 kilometers. So I used a sigmoidal function to assign a decreasing suitability the further from a Skytrain Station.
5)Bus Stop Distance Factor
This image is similar to that of the Skytrain Stations, but because bus service is not as fast as the Skytrain, fewer people would be willing to use them. Consequently, a pub would have to be located very near to a bus stop for that pub to gain any advantage. That's why in the fuzzy image, I used a sigmoidal function to assign a decreasing suitability to a distance of no more than 400 meters, or roughly an 8-10 minute walk.
6)Pub Distance Factor
The idea behind the Pub Distance factor was that a pub wouldn't be viable beside another pub, and that locating too far from an established pub would mean needing to build a new client base. So, there was a suitability range between about 100 meters and 2 kilometers. This was based on the distance image itself, with the closest pubs about 100 meters apart, and the furthest at about 2 kilometers.
7)Sex Ratio Factor
For this image, some data manipulation had to be done. As IDRISI could only do fuzzy analyses on integer data, I had to multiply the original ratios by 1000 in order to achieve some differentiation between the ratios of different DA's. Then, using the fuzzy module,
I used a J-Shaped function to give an equal suitability to all ratios between 0.98 and 1.020, with a narrow range of suitability on either side of the maximum suitability. The reason for this was so that a base of clients would be approximately equal, although it didn't matter too much as people obviously will come from other areas.
8)Income Factor
In this image, I wanted to target areas where individual income after tax was over $7,000 more than 70% of the time. I did this using a J-Shaped Function, giving an equal suitability to all values over 95%. The reason why the income is low is also tied into the next factor.
9)Age Factor
For this factor, I was looking to target areas with a high portion of the population between the ages of 19 and 34. This was because most university students fall into this age bracket, along with young people in the work force looking to unwind. As a result, I arbitrarily chose areas that had at 15% of the population in my age bracket as least suitable. Using a sigmoidal function, suitability increased untill 50%, where all proportions higher than that had an equal suitability.
Once I had set up my factors and constraints, I then had to decide on weights using a Weight module. I had to create a Pairwise Comparison File, which resulted in this figure:
Figure 1. Pairwise Comparison File values.
After I input these values, the module then gave me weights to use for my final analysis. Figure 2 shows the results of running the module:
Figure 2. The eigenvector of weights and consitency ratio
After running the MCE module in IDRISI, I came up with the final output of my Multi-Criteria Evaluation. Additionally, for comparison, I ran the MCE module again, using equal weights this time. Both outputs are shown on the following page.