Scope and topics
Earth Observation (EO) is now indispensable to the environmental and physical sciences, yet its uptake in the social
sciences remains comparatively limited. Social conditions such as poverty, deprivation, segregation, wellbeing, mobility,
access to services have no direct spectral signature; they must be inferred through proxies, models, contextual
knowledge and careful validation. Doing this credibly demands methodological work that sits between remote sensing
and social science: representation learning that captures socially meaningful structure, principled fusion of EO with
surveys, censuses, registers and participatory data, explicit treatment of uncertainty and bias, and research designs that
support inference rather than description alone. It also raises conceptual and ethical questions about what is being
measured, for whom, and with what consequences for the people represented in the data.
This session brings together researchers developing and critically examining EO-based approaches to social scientific
questions across both data-scarce and data-rich settings. Contributions from the Global South, where EO frequently
substitutes for missing statistical infrastructure, are as welcome as work from register-rich contexts such as the Nordic
countries, where EO instead complements and spatially disaggregates existing micro-data. Comparative work, and
studies testing the transferability of approaches between these settings, are explicitly encouraged, as are validation
studies, benchmark comparisons and negative results, which remain under-reported in this literature.
List of topics:
- EO-based measurement of poverty, deprivation, inequality, segregation and living conditions
- Population distribution and dynamics, mobility, migration and forced displacement
- Fusion of EO with censuses, surveys, administrative registers, citizen science and participatory data
- GeoAI for socio-economic inference: representation learning, geospatial foundation models, vision–language
models - Explainability, uncertainty quantification, bias, fairness and validation of EO-derived social indicators
- Causal inference and policy evaluation using EO-derived measures
- Longitudinal and historical social analysis using archival imagery and historical cartographic sources
- EO for public health, service accessibility, environmental justice and exposure to climate risk
- Armed conflict, food security and crisis analytics
- Ethics, privacy, governance, and inequalities in access to EO data, infrastructure and expertise
- Open, reproducible and FAIR workflows; capacity building and training for social scientists
Organizers
Ola Hall Dept. of Human Geography, Lund University, Sweden
Stefanos Georganos, Geomatics, Dept. of Environmental and Life Sciences, Karlstad University, Sweden
Lina Eklund Dept. of Earth and Environmental Sciences, Lund University, Sweden
Adel Daoud ,Institute for Analytical Sociology, Linköping University & Dept. of Computer Science and Engineering, Chalmers University of Technology, Sweden
Lia Antoniou , LF Research Foundation, Sweden
Sai Ganesh Veeravalli, Risk and Environmental Studies, KarlstadUniversity, Sweden