Mapping Deprived Urban Areas Using Open Geospatial Data And Machine Learning In Africa
Multidisciplinary Digital Publishing Institute (MDPI)
Download
Abstract: This study presents a novel methodology for identifying urban deprivation zones in Africa using open geospatial datasets and machine learning models. Focusing on Accra, Lagos, and Nairobi, the approach integrates 50 geospatial indicators?including population density, infrastructure access, flood exposure, informal housing, and socioeconomic risk?into the IDEAMAPS framework. Models trained on three machine learning classifiers achieved over 80% predictive accuracy in mapping deprived urban areas. The authors emphasize the need to complement physical hazard mapping with social deprivation indicators to capture compound vulnerabilities. Ethical design considerations address the risks of stigmatization and eviction when using high-resolution spatial data. The methodology supports scalable, low-cost solutions for urban climate resilience, resource targeting, and informal settlement upgrading. Nairobi's inclusion strengthens Kenya's urban planning capabilities by offering actionable workflows for real-time spatial analysis and adaptive infrastructure allocation in vulnerable neighborhoods.