Tomasz Migdal
Modeling Urban Growth with Existing Urban Built-up Area Within the Toronto Census Metropolitan Area (CMA) using Ordinary Least Squares and Geographically Weighted Regression ©2014
Urban areas are complex systems that are influenced by different processes. Although urban areas provide education, culture, and health care the development and change of urban areas increase congestion, criminality, and pollution. In order to study urban areas, modeling is considered to be an efficient way to understand the factors that influence change. This study applied the ordinary least squares (OLS) and geographically weighted regression (GWR) in order to analyse urban growth through built-up area within the Toronto Census Metropolitan Area (CMA) between 2001 and 2011. The study focuses on census tracts that are not fully developed and conducts a comparison between the OLS and GWR methods. The results indicate that the four independent variables, road network density, distance to central business district, population change, and population density can successfully predict built-up area with Toronto CMA. Through a comparison of the goodness of fit, residuals, and spatial autocorrelation the GWR proves to be the better model over the OLS, indicating that the model can be used to study urban growth through built-up area. Analysis also indicate that GWR can be successfully used to predict future built up area, with relatively high accuracy.