Machine learning for analysis of wealth in cities: A spatial-empirical examination of wealth in Toronto

Abstract

Wealth in the Greater Toronto Area (GTA) continues to grow each year as Toronto's consumer market and population increase. Using a machine learning segmentation based on self-organizing maps, this paper examines the demographics, socioeconomics, and expenditure consumption patterns of the GTA's consumers. The results suggest that SOM may contribute to efficient spatial delimitation tools, enhancing the spatial patterns of clusters in the city of Toronto. The relation to urban areas displays locational neighbourhood characteristics, where the accumulation of wealth is present, pointing out a striking spatial-morphological division between census regions and geographical distribution of wealth in Toronto. In this sense, concerning regional and urban habitats, SOM position themselves as promising tools to measure wealth within highly dense urban cores with significant demographic diversity. While cities that have witnessed rapid urbanization and population growth, such as Toronto, may benefit from integrative methods that use machine learning and spatial analysis to monitor regional and urban disparities.