GeoAI Framework for Assessing Climate Vulnerability in African Urban Transport

Rapid urbanization and increasing climate variability are intensifying vulnerabilities in urban transport systems across African cities, threatening mobility continuity and infrastructure reliability. Conventional GIS-based overlay approaches support hazard visualization but remain limited in capturing the nonlinear interactions and network interdependencies that drive systemic vulnerability. This study develops a GeoAI-based spatial inference framework that reconceptualizes transport climate vulnerability as a probabilistic, network- embedded process rather than an additive spatial representation. The framework integrates environmental exposure, infrastructural sensitivity, and adaptive capacity within a machine-learning architecture that explicitly incorporates transport network topology. The modelling workflow combines data harmonization, network- based feature engineering (including centrality and redundancy metrics), and machine-learning inference to generate continuous vulnerability surfaces. The framework is demonstrated through its application to Lagos, Nairobi, and Cape Town to illustrate its transferability under contrasting climatic and infrastructural conditions. The methodological demonstration indicates that incorporating transport network topology within probabilistic inference enables the consistent identification of structurally critical transport corridors across contrasting urban contexts. By formalizing vulnerability assessment as an inference-driven, network-aware process, the study provides a transferable GeoAI methodology for climate-resilient transport planning in data-constrained urban environments. Housing Quality Research Methodologies in Nigeria: