Submissions Accepted for Presentation at the World Bank Land Conference 2024

The worldwide expansion of human settlements is a major driver of global change and in particular the sprawl of cities poses great challenges. Within just a few years, megacities are sprouting up, urban agglomerations sprawl into hazard zones and entire landscapes are fragmented and irretrievably transformed. But how can the opportunities offered by prospering settlements be put to good use? And how can risk exposure and negative environmental and societal impacts be mitigated? These are key challenges for land governance. And it is here where satellite-based earth observation and machine learning can make important contributions. This article presents the World Settlement Footprint (WSF), digital maps derived from "Big Earth Data", as a valuable tool for information gathering and decision support. The information and knowledge provided by the WSF can help planning, science, policy and business to enable sustainable and informed governance practices and land use strategies.

This paper provides a detailed descriptive analysis of patterns of urbanization across Sub-Saharan Africa. Despite the rapidity of Sub-Saharan Africa's urbanization, little is known about the anatomy of patterns of urbanization across the region due to a lack of detailed and accurate official data on urban populations. To address this gap, this paper applies the “dartboard” algorithm to high-resolution gridded population data for the region, which is derived from digitized maps of the footprints of all buildings in the region. This allows for a consistent definition of urban areas across countries, overcoming the measurement problems that arise from relying on official definitions of urban areas. Using this definition, the paper presents evidence on key empirical regularities that relate to disparities across the urban hierarchies, as well as on the internal structures of cities. Analysis of how the derived urbanization characteristics relate to country characteristics is also presented.

The lack of comparable urban definitions across countries has presented a significant challenge in effectively addressing poverty in both urban and rural areas. This study aims to tackle this issue by comparing subnational poverty statistics across countries, integrating internationally consistent definitions of urban areas into the World Bank’s official global poverty measurement framework. Focusing primarily on 16 Sub-Saharan African countries, the analysis reveals that poverty rates tend to be lower in densely populated urban areas. However, the findings also highlight that urban areas have a higher concentration of impoverished populations than previously estimated. These results underscore the importance of employing consistent urban definitions in cross-country poverty analysis and call for a reevaluation of geographically targeted policies to expedite poverty reduction efforts.

Understanding economic development, land rights and management, and structural transformation requires accurate and high-resolution measurements of poverty and growth. Fine-grained estimates of living standards are also critical to effectively target policies and evaluate development interventions. However, most low-income countries lack the resources and administrative capacity to regularly collect household-level socioeconomic information. Research in the past few years has developed methods for poverty measurement at the village, neighborhood, and satellite tile level based on machine learning (ML) and satellite imagery. This project advances the literature by producing and evaluating the first national-scale satellite-based poverty estimates at the household level, using two large field surveys in Togo and Bangladesh. We also provide guidance to policymakers and practitioners on which aspects of living standards can be reliably measured from satellite imagery, as well as the cost-benefit tradeoffs in choosing appropriate machine learning techniques for such prediction tasks.