DynEO4SLUMS | Sai Ganesh Veeravalli

Earth Observation for Mapping Margins

With the global slum/informal settlement population now estimated to exceed 1 billion — and a significantly more acute situation in low- and middle-income countries (LMICs) — there is a pressing need to systematically monitor the spatial and temporal dynamics of such deprived neighbourhoods and assess the exposure of their vulnerable communities to multiple hazards. Despite its significance, this remains a substantially under-researched domain.

The primary scientific objective of this two-year research project is to exploit the capabilities of Earth Observation (EO) and Artificial Intelligence (GeoAI) to develop methodologies for characterizing the evolution of urban settlements with informal morphologies, assessing their population dynamics, and quantifying multi-hazard exposure over time.

Objectives

The general scientific objective of this research project is to leverage the potential of EO and GeoAI and to develop new scalable methods to monitor and improve our understanding of the spatial evolution of urban settlements with informal morphologies, their population, and exposure to multiple hazards through time, as these are largely under-researched topics.

Expected scientific results:

Innovative EO approaches for deriving variables related to the space-time dynamics of urban settlements with informal morphologies, their population, and exposure to hazards
Insight into the potential of several super-resolution models for a real-life application in a complex urban context
Insight into the potential of models for capturing the space-time changes of complex urban forms.
Advancing the population modelling field beyond the state-of-the-art by tackling some of the most challenging modelling scenarios, harnessing recent advances in DL
Framework for producing consistent indicators that can help monitor progress towards SDG targets

Technical Approach

DEPRIMAP combines Earth Observation data with machine learning to analyze urban environments at neighborhood scale. The framework integrates globally available datasets on buildings, roads, land cover, and derives spatial indicators that capture urban form and infrastructure conditions. These features are used within supervised learning models to identify patterns associated with urban deprivation. Designed for scalability, the approach can be applied consistently across thousands of cities, enabling comparable assessments in data-scarce regions.

Team

Research Outputs

2025

Conf. Proceedings
Understanding Informal Settlement Transformation through Google’s 2.5D Dataset and Street View based Validation

Sai Ganesh Veeravalli, Jan Haas, John Friesen, and 1 more author

In The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2025

Abs DOI Bib Code

Monitoring change in informal settlements remains a critical challenge, particularly in data-scarce contexts across the Global South. While satellite remote sensing provides strong temporal coverage, conventional approaches for mapping the built environment often rely on very high-resolution imagery or LiDAR, which lack consistent temporal availability and are costly to scale especially for capturing vertical growth. This study leverages Google’s Open Buildings 2.5D Temporal Dataset (2016-2023), which offers annual estimates of building presence, count, and height, to detect structural change in Nairobi, Kenya. By analysing differences in building count and average height across 100-meter grid cells, we developed a rule-based framework to identify four key transformation types: vertical densification, horizontal densification, combined densification (increase in both count and height), and decline. To our knowledge, this is the first study to use this dataset to assess vertical change within informal settlements. Validation was conducted through a two-source approach using historical satellite imagery (Google Earth Pro) and archival street-level imagery (Google Street View). A total of 154 grid cells across 13 slum areas were manually assessed, yielding an overall accuracy of 96.75%. Horizontal and combined densification showed perfect agreement, while vertical densification and decline categories had over 80% accuracy. Spatial analysis across slums, adjacent buffer areas, and the broader city revealed horizontal densification as the dominant trend within informal settlements, while vertical and combined growth were more prominent in surrounding zones. These results demonstrate the potential of Google’s 2.5D dataset for scalable, interpretable urban monitoring in rapidly changing environments.
@inproceedings{veeravalli2025understanding, title = {Understanding Informal Settlement Transformation through {Google's} 2.5D Dataset and Street View based Validation}, author = {Veeravalli, Sai Ganesh and Haas, Jan and Friesen, John and Georganos, Stefanos}, booktitle = {The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences}, volume = {48}, pages = {245--251}, year = {2025}, doi = {10.5194/isprs-archives-XLVIII-M-7-2025-245-2025}, url = {https://isprs-archives.copernicus.org/articles/XLVIII-M-7-2025/245/2025/}, projects = {deprimap, dyneo4slums}, }

Funder and partners information