Where O Where Do I Open a New Consignment Store in Vancouver?




Spatial Analysis


To do the spatial analysis I used the Decision Wizard and Weighted Linear Combinations.  I used the Boolean image of commercial areas as my only contraint (link).  My factors were the existence of other stores and demographics, both grouped into neighbourhoods, and the cost of commercial space.  I then combined these in 5 different ways to see how different standardizations, control points, and weights would affect my outcome. All 5 combinations used the same input images: Boolean image of commercial areas, Neighbourhoods with # of stores as thier value, Boolean image of all neighbourhoods with a greater proportion of females aged 15-34, and the interpolated lease cost image for the commercial areas.

Combo #1:
Existing stores: used a monotonically increasing J-shaped curve, so that areas with more stores were seen as much more suitable.
Demographic: used a linear increasing curve.
Lease cost:   used a sigmoidally decreasing curve, with $20/sqft and $50/sqft as the upper and control points, respectively, of the curve.
I then combined these with equal weights.  See the map here.

Combo #2:
This combination used the same control points and standardization curves as #1, but I used the Weight module and the Analytical Hierarchy Process. This utilized pair-wise comparisons between the factors, and the weights and eigenvectors used are:
Existing stores: 0.4200
Lease Cost:     0.5109
Demographic:  0.0691
I chose these values because lease cost is the most important factor, and demographic is the least important. As was seen on the demographic maps, there is a fairly even distribution of men and women within this age group across Vancouver.  And in the areas where there were more men, like Downtown and Strathcona, I believe if there were other stores around the women would come to shop.  Therefore, rent and nearness to other stores is very important, with rent being most important. See the resulting map here.

Combo #3:
Existing stores: used a symmetrical curve, with control points of 0, 3, 6, 10. This means that neighbourhoods with between 3 and 6 stores are very suitable, while the degree of suitibility drops off as it approaches 0 or 10. This tries to take into account the competition factor when there are many of the same type of store.
Demographic and Lease cost parameters were the same as for #1 and #2.
I used equal weights for all 3 factors. See the map here.

Combo #4:
This combination used the same standardizations as #3 and the same weights as #2.
See the map.

Combo #5:
For this combination I only changed the control points for the Lease Cost factor. For the previous 4 combinations the point at which suitability began to drop off was at $20/sqft, and the point at which suitibility began to level off toward 0 was at $50/sqft. For #5 I changed this so that suitibility began to drop off at $15/sqft and level off toward 0 at $30/sqft.  I felt this would take rent costs into even greater consideration, and would hopefully come up with areas that would be ideal to open a store and still be completely affordable.  The other standardizations and all the weights are identical to Combo #4. See the map.



Go back to the Introduction, the Methodology, Data Acquisition, or ahead to the Conclusion or the Problems and Errors.