Neil Ossher

Population projections are commonly used in both the private and public sector for strategic decision-making. In the private sector, businesses often use projections to forecast consumer demand for site selection purposes. As a result, there is an increasing need for population projections at smaller spatial scales such as neighbourhoods, census tracts and dissemination areas. Accuracy of population projections in these small areas is affected by many factors including residential development, changing boundaries, and data availability. This paper attempts to improve the accuracy of existing population projection models using ancillary data on residential developments using the municipalities of Markham and Richmond Hill as a laboratory. Building records and development applications are used to adjust existing projection models and produce population projections for 2006, 2011 and 2016. The 2006 projections are compared to the actual census counts to evaluate the models' performance. The adjusted linear extrapolation model is found to be the most accurate, whereas some bias is detected in the logistic growth model. The adjusted linear model is then used in a case study of Chinese retail within the study area. In conclusion, the results of this are generally mixed with some models showing improved accuracy and others displaying decreased accuracy. This analysis provides valuable insight for businesses because it develops a method for adjusting existing models and demonstrates how they can be applied