Author(s): Vaz, Eric ; Damásio, Bruno ; Bação, Fernando ; Shaker, Richard Ross ; Penfound, Elissa
Date: 2023
Persistent ID: http://hdl.handle.net/10362/154833
Origin: Repositório Institucional da UNL
Subject(s): Urban Studies; SDG 2 - Zero Hunger; SDG 3 - Good Health and Well-being
Vaz, E., Damásio, B., Bação, F., Shaker, R. R., & Penfound, E. (2023). Urban habitats and food insecurity: lessons learned throughout a pandemic. Habitat International, 135, 1-11. [102779]. https://doi.org/10.1016/j.habitatint.2023.102779
Background An increasing amount of literature raises the issue of food deserts and urban heterogeneity in larger metropolitan cores throughout North America. Specific to Canadian cities, the disparity between access to health, education, and affordable food is of growing concern. Recently, these drivers seem to be significantly linked to the propagation of COVID-19. This paper explores the spatially-explicit dynamics of food deserts in Toronto, by integrating Geographic Information Systems and machine learning to understand the clusters of food deserts. The integration of spatial analysis with self-organizing maps (SOM) offers insights on the relation between neighborhoods, geodemographic profiles and urban characteristics, and whether one might expect consequences of food insecurity given COVID-19. Methods The paper starts out with developing a machine learning algorithm based on SOM to define meaningful clusters within the hedonic dataset. Further to this, an exploratory regression was built per cluster as to allow an exploratory spatial analysis to derive an explanatory framework for the key characteristics of socio-economic profiles within the Greater Toronto Area and impacts of SARS-CoV-2. Results The findings suggest that there are clear spatial profiles within the urban core of Toronto in regards to food deserts, showing a direct relation between socioeconomic characteristics and the results on environmental injustice and livability. These profiles are strongly linked with the areas of COVID-19 occurrence, and share a very similar socio-demographic profile, particularly in regards to young and lower income families. Conclusion There are several food deserts currently in Toronto, Ontario. The integration of policies that involve public health and spatial decision-support, particularly when linked to machine learning to aggregate characteristics of big data, establish a multi-functional understanding of the complexity of food security. This has a direct relation with diet, environment, and the opportunity to enhance subjective well-being in city cores.
