Global Mapping of Informal Settlements using Satellite Imagery and Open Datasets

Informal settlements, where groups of asylum-seekers, refugees, or internally displaced people settle in self-identified, spontaneous locations, often lack basic services and are highly vulnerable to climate-driven risks. Approximately 1 billion people currently live in informal settlements, driving significant research attention to issues such as property rights and governance strategies. However, the lack of large-scale location and boundary data for informal settlements poses a major obstacle to conducting in-depth quantitative analyses. Existing approaches to identifying informal settlements have been constrained to individual cities or districts due to their reliance on costly imagery, field survey, and localized expertise. To address this gap, we develop a novel, global framework to identify permanent informal settlements using satellite imagery and publicly available datasets. First, building footprint and nighttime light datasets were combined to isolate human settlements with low energy availability, effectively narrowing the focus to potential settlement areas and significantly reducing computational costs. Next, leveraging high-resolution Sentinel-2 imagery, we estimated the similarity of each pixel's spectral principal components to those of high-confidence samples, allowing us to distinguish informal settlements from broader human settlement areas. Finally, nearly 10,000 informal settlement points provided by the United Nations were used to refine our results, ensuring that only officially recognized informal settlements were retained. This fast and scalable framework overcomes the data scarcity challenges typically faced in resource-poor regions, where informal settlements are most prevalent. The results of this study provide an unprecedented foundation for deeper quantitative analyses of informal settlement areas, supporting global efforts to achieve the Sustainable Development Goals and address urban vulnerabilities.