SATELLITE AND ARTIFICIAL INTELLIGENCE IN MAPPING MULTIDIMENSIONAL POVERTY IN AFRICA

Context and background Multidimensional Poverty (MP) considers poverty in multiple dimensions of deprivations such as health, education, energy, the standard of living and access to basic services. MP remains a major challenge in Africa, with a large proportion of the population living in MP. According to United Nations Development Programme (UNDP), Africa has shown the highest Multidimensional Poverty Index (MPI) having over 40% of its population living in MP. Goal and Objectives: This paper is a review, aimed at assessing the potential of the integration of satellite and Artificial Intelligence (AI) in mapping MP, with a specific focus on Africa. Methodology: Based on the reviews of past studies, the combination of satellite data such as nighttime light, daytime satellite imagery and high-resolution settlement data in combination with techniques such as field surveys, statistical correlation models (transfer learning) and AI (deep learning) has been applied in mapping MP. Results: The findings from studies show that the combination of satellite data and AI has the capability of providing more accurate and granular MP maps, compared to the traditional approach. Again, this paper explains the concept of MP with a specific focus on Africa and presents a map depicting the current MPI in African countries. Finally, pitfalls especially in the accuracy, granularity and frequency of MP data were identified. Consequently, the satellite and AI approaches are recommended for more accurate, frequent, cost-effective and granular data, required in mapping poverty and design of interventions that effectively address the needs of the vulnerable populations in Africa.