The previous research team tried many different models based
upon built form to classify suburban neighbourhoods. A
pilot study of the Ottawa-Gatineau CMA was used to test various
models, with the results checked by Google Earth and Google Street
'Inner City' neighbourhoods were identified using Professor David Ley's definition (2006). Census tracts that had higher than the metropolitan area's population of houses built before 1946. This method had been used in other studies, but pre-1946 buildings have become less reliable method to identify inner city neighbourhoods as new construction occurs and brownfield redevelopment has been completed in large Canadian cities.
'Rural' neighbourhoods are classified as census tracts with a population density of less than 105 people per square kilometre based upon a previous research study (Ohio University, 2002). This has subsequently been updated to a Statistics Canada article for the Transportation Method and Density Method Rural (Exurban) definitions.
The remaining suburban neighbourhoods were identified by a combination of age of housing, proportion of singles, semi and attached units, and home ownerships ratios.
We produced a 2006 classification that matched the Ottawa air photos interpretation, but this definition was not useful for earlier census years or for other cities. Many Canadian cities have apartments or townhouses in their outer suburbs, such as Kanata or Orleans in Ottawa. These units confound a classification system based upon built form. A built form definition that gave reasonable results in Ottawa might produce a classification that made little sense in other cities.
Here are some examples: