Application of Neural Networks to Optimize Construction of Water-Energy Datasets for Sub-Sahara Africa

In January 2025, data collectors for the United Nations Sustainable Development Group estimated that about 600 million Africans lacked access to electricity. [1] In March of 2024, UNICEF and the World Bank suggested that nearly 319 million sub-Saharan Africans could not access reliable drinking water. [2], [3] Energy and water are inextricably linked. [4] However, researchers persist in treating these topics as mutually exclusive. To advance concepts and opportunities offered by renewable energy resources like solar, wind, and biomass energy, this paper proposes using neural networks concepts to bridge the gap in data reporting. While the water-energy nexus is important to encourage investments, production, and economic viability, we did discover that the best water-for-energy model has three hidden layers and the energy requiring water model as two hidden layers. The output of the network was bilateral for each resource and can be used when designing studies or regulating datasets.