Leveraging Machine Learning and Earth Observation for Mapping deprived areas in the Global South

Informal settlements are home to 1.1 billion people, and represent a growing share of the urban population. Despite their central role in the global housing crisis, many settlements and their dynamics are still insufficiently captured in official statistics and planning processes and pose significant challenges for public entities in meeting the needs and priorities of residents and communities. Data gaps, such as missing variables in surveys, limited spatial and temporal granularity, and a lack of validated information, constrain the monitoring of SDG 11.1 and the ability of governments, communities, and development partners to design and monitor effective upgrading and housing interventions. Building on ongoing work, this session brings together researchers, civil society, development partners, and national and local governments to reflect on how Geospatial and AI (GeoAI) technologies can support the transformation of informal settlements when embedded in inclusive and ethical data ecosystems. Academic partners from ETH Zürich, the University of Twente, and HafenCity University Hamburg will present recent research and advances in mapping and monitoring informal settlements using GeoAI combined with local (e.g., community-based) data sources. Representatives from community-based organisations and networks such as SDI and HOT will provide perspectives on community-led data, while local government representatives (e.g. from Brazil and Nigeria) will focus on governance and policy uptake. Thus, the session will showcase recent advances in GeoAI-based approaches for mapping and monitoring informal settlements, and the opportunities they present while critically reflecting on their limitations and risks. Particular emphasis will be placed on the co-production of data, combining GeoAI-based approaches with citizen-generated data and local knowledge to improve data quality, trust, and relevance. Through concrete case examples, the session will explore how spatial data can inform slum upgrading, housing investment prioritisation, service provision, and the prevention of forced evictions by highlighting alternative pathways to improve resilience and liveability in informal settlements. A moderated dialogue format will enable exchange between researchers, community representatives, and public authorities on how data-driven approaches can move beyond visibility towards tangible improvements in living conditions.