Hidden and fragmented patterns of morphological informality across African cities revealed by open geospatial data

Abstract Informal settlements are among the fastest-growing forms of urbanization, yet their spatial distributions within cities remain under-represented and inconsistently documented in official statistics and global urban datasets. Conventional assessments often identify large, contiguous informal neighborhoods but miss smaller and fragmented forms embedded within otherwise formal urban areas. This spatial invisibility limits the ability of cities to identify where informality intersects with infrastructure deficits, environmental risk and service provision. Here, we develop a scalable framework to map morphological informality using open satellite imagery and global building datasets. City-specific machine learning models are combined using a similarity-weighted ensemble that enables cross-city transfer while retaining local sensitivity. Applying the approach across 84 African cities, we generate 10 m resolution probability surfaces describing the likelihood of morphologically informal settlements. The resulting maps reveal substantial intra-urban heterogeneity in the spatial expression of morphological informality. In addition to identifying extensive informal settlements, the framework captures fragmented clusters embedded within more formal neighborhoods, which are patterns frequently overlooked by existing global mapping products and coarse statistical indicators. By revealing these hidden spatial structures, the approach provides a harmonised and scalable workflow with global applicability, enabling rapid updates to monitor change and supporting the integration of contextual, place-based knowledge into assessments of informal settlement status and impacts.