In the sprawling metropolitan areas of Africa, peri-urbanization is a phenomenon that’s as complex as it is rapid. This unplanned urban growth presents significant challenges for land management and sustainable territorial planning. Now, a groundbreaking study published in *Urban Science* (translated from French) offers a new approach to detecting and analyzing these dynamic peri-urban areas, with potential implications for the energy sector and beyond.
Led by Asmaa Moussaoui from the Department of Cartography and Photogrammetry at the School of Geomatics and Surveying Engineering, part of the Agriculture and Veterinary Medicine Institute Hassan II in Rabat, Morocco, the research employs a deep learning model to map and monitor peri-urban territorial dynamics. The model, trained on a manually annotated dataset covering the Casablanca metropolitan region, classifies areas into four categories: urban, peri-urban, rural, and water.
The study uses the Global Human Settlement Layer and Global Land Analysis and Discovery Land Cover data to feed the Multi-Layer Perceptron model. “This model achieved high accuracy, around 90.6%,” Moussaoui explains. “It performed exceptionally well in identifying urban and rural areas, but peri-urban areas presented some challenges due to their transitional land patterns.”
The diachronic analysis conducted from 2005 to 2025 revealed a significant expansion of peri-urban areas, with an increase of approximately 28,000 hectares at the expense of rural lands. These findings could be a game-changer for policymakers, offering valuable insights for sustainable land-use planning and anticipating urban sprawl dynamics.
For the energy sector, understanding these territorial dynamics is crucial. As cities expand, so does the demand for energy infrastructure. Accurate mapping of peri-urban areas can help energy companies strategically plan the development of new power plants, distribution networks, and renewable energy projects. It can also aid in identifying areas at risk of energy poverty, ensuring that infrastructure development keeps pace with urban growth.
Moreover, the study’s use of deep learning and geospatial data sets a new standard for territorial analysis. As Moussaoui notes, “The Shapley Additive Explanations method ensured model interpretability, making our findings more accessible and actionable for policymakers.”
The research not only sheds light on the complexities of peri-urbanization but also provides a robust tool for managing this dynamic process. By offering a clear, data-driven approach to territorial planning, it paves the way for more sustainable and resilient urban development. As the energy sector continues to evolve, such tools will be invaluable in ensuring that infrastructure development aligns with the realities of urban growth.
In the words of Moussaoui, “Our work offers a new perspective on peri-urbanization, one that could help shape the future of our cities and the energy systems that power them.” With its high accuracy and potential for commercial impact, this research is a significant step forward in the field of territorial planning and urban development.