Aside from the numerous small mistakes that I brought upon myself throughout this project, there were some definite areas where this project could have been improved. Although I do consider the results of the analysis useful, perhaps newer data would have made the analysis more indicative of current and future trends. One of the main problems with this project is that it is looking forward at least 6-8 years. So many new developments could occur by then which would have an impact on the spatial question of where to locate this facility. In terms of data collection, actual land value data would have been more accurate than housing cost per census tract. Also, for ice rink data, it would have been useful to build in information about each of the 38 existing facilities and creating the Thiessen polygons based on weighted factors accordring to rinks, or square meters of ice surface. I considered doing this but the background research would have been too great considering there is no data available on this. The idea of considering existing ice surface facilities is not that the Olympic facility would be in direct competition with a local community rink, but in other areas where these facilities have been built (Calgary and most recently in Salt Lake City) they have been used extensively by the local community after the Olympics.

A methodolgical change to this analysis might be to leave the landuse data as a constraint in the MCE analysis, however I have justified earlier why this was not done. But it might allow the other factors to trade off more with each other if the weight given to landuse was distributed among the other factors. It would also be interesting to perform to completely two separate MCE analysis' , one with social factors and the other with conventional factors to see how the two perspectives on site location compare.

Any type of market analysis data is almost impossible with this problem. It makes it much more difficult than a business site location analysis. The amount of people involved in speed skating, or ice sports in general would definitely be hard to acquire.

Also in each layer where I digitized data which was not otherwise available, I used pure hand - eye coordination to locate the points. This type of digitzing is prone to locational errors and its likely that they entered the data sets. However, the impact that these types of errors would have on the outcome of the analysis is minimal.