Leveraging Machine Learning and Street View Imagery to Study the association between social inequalities and visual walkability in the South Africa Context

Walkability is crucial for improving urban sustainability. Most studies on walkability focused in developed countries, while only a few studies have been conducted in the Global South due partially to limited data availability. To address the limitation, this research calculated visual walkability using Street View Imagery and computer vision techniques in Johannesburg, South Africa, and then explored the relationship between social-economic indicators and walkability. The results offer insights on the social and spatial inequality of walkability in the South African context.