High-resolution wealth maps reveal Africa’s nonlinear trajectory of development and inequality

High-resolution wealth maps reveal Africa’s nonlinear trajectory of development and inequality | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article High-resolution wealth maps reveal Africa’s nonlinear trajectory of development and inequality Qingyan Meng, Xin Tan, Qikang Zhao, Linlin Zhang, Liqiang Zhang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8234710/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Economic well-being data are critical for understanding development conditions and monitoring poverty, yet micro-level socioeconomic conditions in Africa remain poorly understood. Here, we present a high-resolution wealth mapping framework that integrates household survey indicators with spatial wealth proxies, producing the first 10-m economic well-being maps across 30 African countries. Our model explains 74% of the variance (R²=0.74) and our results reveal that ~ 36% of the population—about 340 million people—still reside in the least-developed areas. Expanded analysis of nearly 60,000 settlements further shows that Africa remains in a state of “equilibrated poverty”: while economic growth tends to widen internal disparities, development trajectories follow a nonlinear U-shaped pattern rather than a monotonic trend. Under extreme poverty, inequality actually intensifies, creating compounded risks of poverty and social exclusion. These findings uncover spatial patterns invisible to coarse-scale studies and provide new evidence for poverty alleviation and equitable infrastructure planning across the Global South. Earth and environmental sciences/Environmental social sciences/Sustainability Scientific community and society/Geography Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Social and economic well-being is a fundamental reference for understanding development and guiding policies. Globally, poverty and inequality are tightly intertwined: economic growth has not provided equal opportunities, instead, in many countries, the urban-rural divide and regional disparities have been further widened. Underdeveloped regions face not only poverty but also multiple deficiencies in infrastructure, services, and basic living conditions [ 1 ]. Such multidimensional deprivation often overlaps with inequality, creating a vicious cycle of poverty and social exclusion. Recent studies [ 2 ] suggest that, even under optimistic scenarios, the global goal of poverty eradication will be delayed by more than two decades, underscoring the urgency for intensified efforts. However, there is still little clarity on how many people live in underdeveloped regions and how poverty and inequality interact[ 3 ]. This knowledge gap limits accurate assessments of development conditions, weakens policy design, and hinders the achievement of the Sustainable Development Goals (SDGs)[ 4 ]. The Demographic and Health Surveys (DHS) Program has conducted more than 400 surveys across over 90 countries, providing accurate and representative data on population, health, HIV, and nutrition ( https://www.dhsprogram.com/ ). While these surveys help fill critical data gaps, their infrequent updates—often three years apart within the same country—limit their ability to capture dynamic socioeconomic conditions across Africa [ 5 ]. To address this limitation, recent studies have sought to integrate DHS data with Satellite-observation methods characterized by rapid updates and broad coverage, producing large-scale socioeconomic maps such as household asset wealth [ 5 ], access to improved housing [ 6 ], child development [ 7 ], the impacts of solid cooking fuel use [ 8 ], educational attainment [ 9 ], and poverty rates [ 10 ]. These efforts represent important advances toward spatializing well-being, yet most are still restricted to the village scale (Table 1 ), limiting their ability to reveal intra-community disparities or to support localized decision-making. Table 1 An Overview of current map products related to poverty or economic well-being. Map Product Description Coverage Duration Source Limitation Global Subnational Atlas of Poverty (GSAP) Subnational poverty estimates Global 2010, 2019, 2021 WorldBank Lack of ability to analyze further Improved Housing Map Proportion of households living in improved housing Sub-Saharan Africa 2000, 2015 Tusting et al.[ 6 ] 5 km gridded data; limited applicability in community-level planning Informal Atlas (AoI) Vector maps of informal settlements 147 cities, 102 countries selected years from 2000 to 2020 Samper et al.[ 21 ] Crowdsourced data; partial representation of informal areas Pilot high resolution poverty maps Gridded estimates of poverty population 4 countries 2013 Tatem et al. [ 10 ] Early-stage poverty mapping initiative High-resolution poverty maps in Sub-Saharan Africa Asset-based wealth index 44 Sub-Saharan African Countries 2022 Lee et al. [ 22 ] At the settlement level Mapping urban slums and their inequality in sub-Saharan Africa 100m Slum Mapping and Wealth Index 32 Sub-Saharan African Countries 2025 Li et al. [ 43 ] Only in urban areas Most low- and middle-income countries (LMICs) are undergoing or will soon experience rapid urbanization, accompanied by the unexpected expansion of slum-like communities[ 11 ]. To support the transformation of these areas into "inclusive, safe, resilient, and sustainable" communities according to the SDG11, understanding residents' economic and living conditions is an urgent priority. Household Asset Index (HAI) widely recognized as one of the core indicators for measuring well-being, reflecting an individual's satisfaction with their material conditions, encompassing all aspects of income, assets, access to basic services[ 12 ][ 13 ]. The index is specifically constructed based on possession of physical assets (e.g., housing, vehicles, durable goods), housing conditions (e.g., water supply, electricity, sanitation), and essential utilities. Due to its independence from direct monetary income data, the HAI is particularly effective in regions where income is difficult to measure or household-level data are scarce. These advantages have contributed to its widespread adoption in socioeconomic surveys across many African countries[ 14 ][ 15 ][ 16 ]. Internationally, multiple methods have been proposed to measure the HAI, including DHS Wealth Index [ 17 ], Human Development Index [ 18 ], Multidimensional Poverty Index [ 19 ], and International Wealth Index (IWI) [ 20 ]. Among them, the IWI stands out for adjusting cross-country differences in survey design and asset variables, allowing for comparable assessments at both regional and global levels. Recent studies[ 22 ] further demonstrate that IWI, when integrated with Earth observation and micro-survey data, provides a more accurate representation of household wealth distribution, especially in Africa. Building on this evidence, we adopt IWI as the target indicator for high-resolution mapping. To gain deeper insights into Africa’s current micro-level socioeconomic conditions, we extend indicators traditionally constrained to the settlement scale onto high-resolution maps. Specifically, drawing on approaches from fine-scale population mapping researches [ 23 ][ 24 ][ 25 ], we develop a deep learning framework that combines regional survey indicators with fine-grained spatial proxies: the village-level (~ 5km) IWI scores, constructed from DHS data, are projected onto high-resolution (10-m) maps under the guidance of spatial wealth proxies. These proxies are composite indicators constructed from 10-m features of housing quality, ecological environment, and socioeconomic activity. Applying this framework, we produced the first 10-m wealth maps for 30 African countries, providing unprecedented fine-grained insights into population distributions across different levels of development. By aggregating these fine-grained development conditions, analysis of more than 60,000 settlements further reveals marked regional disparities and a nonlinear U-shaped pattern between development and inequality. These patterns not only help distinguish divergent national development pathways but also underscore the compounded risks of poverty and social exclusion under extreme deprivation. Overall, our findings uncover hidden dynamics of development and inequality in Africa and provide a new foundation for poverty alleviation and equitable infrastructure planning. 2. Result 2.1 Performance and Accuracy Analysis To validate the accuracy of the proposed model, we conducted a multi-scale evaluation across 30 countries in sub-Saharan Africa. The assessment metrics are based on the average results of five-fold cross-validation. The model demonstrates strong predictive performance for the village-level IWI derived from DHS data (Fig. 1 c), achieving an overall coefficient of determination (R²) of 0.74 and a mean absolute error (MAE) of 0.24, indicating the effectiveness in estimating wealth across countries. Overall, we find that predictions are more stable in rural areas than in urban samples (Fig. 1 a). In several countries, such as Burundi (BDI), Zambia (ZMB), and Malawi (MWI), due to the complex spatial heterogeneity of urban environments, urban prediction errors are noticeably higher. Spatially (Fig. 1 b), accuracy is significantly higher in West Africa compared to East and Southern Africa with broad national coverage, including six countries achieved an R² above 0.8, and two-thirds of the countries exceeded an R² of 0.7. In contrast, South Africa (ZAF) recorded the lowest accuracy, with an R² of 0.47. We further validated our method by comparing it against leading approaches in spatial wealth prediction. The results show high consistency with the cross-country models proposed by Lee et al. [ 22 ] and Yeh et al. [ 5 ], both yielding R² values above 0.75. 2.2 Wealth Distribution and Inequality in Africa To reveal Africa’s overall socioeconomic conditions, we mapped the distribution of development level (mean wealth) and internal inequality (Gini coefficient) across nearly 60,000 densely populated settlements, each representing a two-kilometer situation. The results show a significant association between development level and internal inequality (Fig. 2 b, r = 0.59), coupled with a characteristic nonlinear U-shaped curve. Notably, the segmented regression identified two critical inflection points (Fig. 2 b, c 1 = − 0.66, c2 = − 0.38), with about 63% of settlements concentrated at the “bottom of the ladle” (c1 < wealth < c2), characterized by low wealth and low inequality—a state of “equilibrated poverty”. Beyond this range, settlements diverge along two trajectories: one follows a “high-wealth–high-inequality” pathway (wealth > c2), where urbanization drives rapid growth alongside widening disparities; the other falls into a “low-wealth–high-inequality” state (wealth < c1), marked by the simultaneous intensification of poverty and inequality. Our results also reveal evident regional disparities (Fig. 2 c, 2 d). Specifically, North Africa accounts for only 7% of settlements. Given their concentrated distribution along the coastal economic belt, it exhibits the highest average wealth and inequality, with 43% of settlements falling within the top 20% of the Africa wealth distribution. In contrast, West Africa contains more than half of all settlements (54%). Due to its widespread rural settlements, it exhibits the lowest average wealth and about 27% of settlements in the bottom 20%, reflecting widespread poverty. At the national scale (Fig. 2 e), quintile-based analysis further identifies countries exhibiting two distinct development patterns (See Figure S1 ). “High-wealth–high-inequality” countries such as Gabon and Rwanda reflect a concentrated development pattern, whereas “low-wealth–high-inequality” countries such as Togo and Gambia exhibit compounded deprivation risks. 2.3 Space Wealth Index map - Angola We present a Spatial Wealth Index map, Angola as an example, to further illustrate development disparities at the national level (Fig. 3 ). Angola's development pattern exhibits a “coastal-to-interior” gradient, with the aggregation curve along the latitude-longitude axis indicating a significant synergistic distribution relationship between wealth and population (r = 0.81, p < 0.01). In the capital Luanda, one-quarter of the population accounts for nearly 60% of the nation's wealth index accumulation (calculation method see Supplementary Note 4), illustrating Angola's highly concentrated development pattern. Central provinces, such as Malanje and Cuando Cubango, exhibit generally moderate levels of development, with wealth concentrated in their core areas. Eastern provinces like Moxico and Alto Zambeze, however, remain in a state of widespread poverty, with reported poverty rates exceeding 50% and reaching over 70% in some regions [ 40 ]. Spatial autocorrelation analysis further indicates that poverty settlements (bottom 20% in Angola’s wealth distribution) are significantly clustered in rural regions of Huambo, Malanje, and Lunda Norte (Moran’s I ≈ 0.41, p < 0.001), where structural constraints—including inadequate infrastructure, sparse population, and complex terrain, which hinder development opportunities[ 41 ]. 2.4 Population Distribution by Wealth Tiers cross 30 African Countries Beyond the Angola case, we further analyzed the proportion of the population across different wealth tiers in 30 African countries (see Supplementary Table S1 for details). The total population of these countries is approximately 940 million, with 36% (about 338 million) residing in the least developed regions (the bottom 20% in wealth distribution). This proportion is consistent with estimates from the World Bank Poverty and Inequality Platform (PIP), which report that around 35% of Sub-Saharan Africa’s population lived below the $ 1.90/day extreme poverty line in 2019 [ 42 ]. This consistency aims to validate the effectiveness of using the lowest wealth quintile as a proxy for poverty and does not imply equivalence to the extreme poverty line. At the national level, countries such as Gambia (72.5%), Senegal (68.7%), Sierra Leone (66.8%), and Togo (64.7%) exhibit poverty rates exceeding 60%, far above the regional average, underscoring the acute vulnerability of West Africa and parts of Southern Africa and the urgent priority for targeted policy interventions. 3. Discussion Although settlement-level maps can reflect basic socioeconomic conditions [ 5 ][ 22 ], high-resolution wealth maps provide additional insights into the internal disparities across different functional zones (Fig. 5 a–d), consistent with objective socioeconomic realities. Notably, informal settlements exhibit substantially lower wealth, demonstrating the map’s ability to detect spatial deprivation in urban environments (Fig. 5 d). Compared to the 100-m wealth maps restricted to major urban areas by Li et al. [ 43 ], our 10-m maps not only provide finer spatial details but also achieve full urban–rural coverage. This makes them an effective and low-cost complement to traditional household surveys, which typically require investments of nearly one billion USD annually across low-income countries to monitor indicators related to the Sustainable Development Goals (SDGs)[ 50 ]. Moreover, wealth maps can be viewed as a comprehensive reflection of socioeconomic development. When combined with population data, they can answer specific questions such as “Where does poverty exist and how massive is it”— spatial and structural patterns invisible in coarse-resolution data [ 5 ][ 22 ]. Importantly, our framework remains robust even under limited data conditions: when trained with only 10% of the samples, the model still performs well (Table S3, test R² = 0.54). Beyond wealth estimation, this framework can be extended to other socioeconomic indicators, including the Multidimensional Poverty Index (MPI), Asset Index Score, household consumption expenditure, and per capita consumption. Poverty in Africa remains extensive, with many countries exhibiting an imbalanced population distribution of “large proportions at both ends and a hollowed middle,” reflecting the dual challenges of current development. By aggregating these fine-grained differences, we found that settlement-level trajectories are more complex than commonly assumed: although economic development tends to widen micro-level disparities, the process is not linear—neither “the richer, the more unequal” nor “the poorer, the more equal.” Instead, it follows a nonlinear U-shaped curve [ 44 ], emphasizing the dominance of a large “equilibrated poverty” group that consistent with previously mentioned conditions of poverty and inequality, while offering new insights into their underlying dynamics. For this “equilibrated poverty” group, policy priorities should focus on enhancing inclusive efficiency, expanding education, basic services, and small and medium enterprise (SME) support—to help populations escape the “poverty trap” and move into middle-wealth levels [ 45 ]. For those still in extreme poverty, the priority lies in improving infrastructure and living conditions to mitigate multidimensional deprivation and avoid compounded cycles of poverty and inequality [ 46 ][ 47 ]. Meanwhile, in rapidly urbanizing communities, wealth accumulation is evident but internal inequality is also intensifying, calling for equitable infrastructure distribution and institutional redistribution to prevent excessive disparities and the growth of informal settlements [ 48 ]. Finally, it is expected to offer a new instrument for aligning macro-level policy goals with micro-level realities. Specifically, they enable cross-validation between national targets (e.g., ~ 60% located in the “ladle bottom”) and micro-level patterns (e.g., the bottom 20% population share), ensuring that poverty alleviation and equity strategies are both evidence-based and geographically precise[ 49 ]. 4. Methods This study proposes a high-resolution spatial wealth estimation framework that integrates multi-source remote sensing imagery with auxiliary data. Overall, we design a deep learning approach jointly supervised by village-level wealth data (5 km × 5 km) and a spatial wealth proxy (10 m × 10 m). The technical framework comprises four main components: A. International Wealth Index (IWI): We construct a standardized wealth index using principal component analysis (PCA) based on asset-related variables from Demographic and Health Surveys (DHS), including five types of durable goods, two public services, and three housing characteristics. This index serves as the supervisory signal for model training and evaluation. B. Spatial Wealth Proxy Features: High-resolution proxy features—including built-up area, building density, NDVI, nighttime lights, road network density, and POI density—are used to construct a "spatial wealth proxy map," providing pixel-level spatial guidance during model training. C. Multi-Source Input Imagery: The model integrates Sentinel-1 radar imagery, Sentinel-2 multispectral data, and SDGSAT-1 nighttime light imagery to enrich spatial context and semantic features, which are fed into a UNet-based architecture. D. Loss Monitoring Strategy: A joint supervision mechanism is established, combining region-level wealth index and the spatial distribution characteristics of the proxy map to guide model optimization. This strategy enhances both predictive accuracy and spatial coherence of the output. DHS Survey Data We collected 51 survey rounds conducted between 2015 and 2024 across 30 African countries. Each country is typically surveyed every 3 to 5 years. For privacy protection, household-level data are aggregated at the cluster level, usually corresponding to a village and each cluster is geo-referenced by GPS coordinates. After preprocessing, approximately 23,700 valid clustered datasets were ultimately formed. It should be noted that to protect respondent privacy, the DHS offsets cluster coordinates by up to 10 kilometers in rural areas and 2 kilometers in urban areas. Although this spatial displacement introduces uncertainty into the mapping result, previous studies [5] have shown that the wealth index remains robust under such displacement, with R² variations less than 0.07 when coordinates are shifted by 2.5 km. Wealth Index Construction Following prior studies [22], we calculated the IWI for each household based on ownership of ten specific assets: five durable goods (television, refrigerator, telephone, bicycle, and car), access to two public services (water and electricity), and three housing characteristics (number of bedrooms, flooring material, and quality of sanitation facilities). The index was constructed using the first principal component derived from a principal component analysis (PCA). The IWI accounts for differences in survey design and asset variables across nations, enabling the construction of globally or regionally consistent wealth maps. Satellite Imagery Google Earth Engine [29] provides archival surface reflectance imagery from Sentinel-2 and Sentinel-1 satellites since 2015. For Sentinel-2, we selected the RGBN bands. Sentinel-1 data were extracted from the Interferometric Wide (IW) mode, including both VV and VH dual-polarization channels. To enhance data quality, we removed pixels with low signal-to-noise ratios (below –30 dB), then computed multi-channel mean composites and applied median filtering for noise reduction [30]. For each DHS cluster, we centered a 5120-meter cropping window on the provided GPS location to cover DHS spatial displacement (typically less than 2 km). This is a widely used preprocessing strategy, although it may introduce some noise, we consider local conditions to be generally stable. Notably, the collected imagery (excluding the nighttime light imagery) was gathered within one year before and after the survey period. It was subsequently processed using median filtering to ensure the highest-quality imagery closest to the survey timeframe. Auxiliary Data and Dataset Partitioning Building footprint data were obtained from Microsoft’s global building footprint product [31], this dataset effectively captures building form and density characteristics across Africa. Additional auxiliary data include road networks and points of interest (POIs) from the Humanitarian OpenStreetMap Team (HOT); Normalized Difference Vegetation Index (NDVI) derived from Sentinel-2 imagery; building area grids from the Global Human Settlement Layer (GHSL) [32]; and 10-meter nighttime light imagery from the SDGSAT-1 satellite, captured in 2021. Although these data were collected at different times, the built environment exhibits high structural stability over the short term (3–5 years) [33] and allows for acceptable tolerance to temporal mismatch in features such as buildings and infrastructure. During dataset construction, Sentinel imagery and auxiliary data were first paired at the sample level. Samples with severe cloud contamination were removed, resulting in a final dataset of 21,233 matched samples. This dataset was partitioned using a five-fold cross-validation strategy. First, five test sets were created, each derived from data collected in at least three independent national surveys and containing 1,800 to 2,300 samples (approximately 10% of the total sample size). None of these test sets were used for training. The remaining data were randomly split into training and validation sets in an 8:2 ratio. Detailed partitioning information is provided in Table S3. Region-Level and Pixel-Level Guided Deep Learning Model We propose a CNN architecture that integrates multi-source remote sensing data. The network is built upon the UNet backbone [34], with a ResNet-50 encoder [35] for feature extraction from Sentinel-1 and Sentinel-2 imagery. Additionally, we design a feature enhancement module, LightNet, to encode SDGSAT-1 data, which is subsequently fused with the main feature stream. The architecture is described as Supplementary Note 1 and Note 2. Spatial Wealth Proxy Construction The IDEAMAPS framework [36] constructs a multidimensional conceptual system for urban poverty, encompassing socioeconomic conditions, infrastructure and services, environmental quality, and unplanned urbanization. While it has demonstrated effectiveness in localized areas such as Nairobi [27], its application at larger scales remains constrained by limited data availability and uneven label distribution. To address this, we focus on extracting indicators related to urban form and socioeconomic status [37][38][39], supporting large-scale spatial modeling across Africa. Specifically, the spatial wealth proxy index was constructed using six high-resolution features—building area, building density, vegetation index, nighttime light intensity, road network density, and point of interest density—to reflect multidimensional characteristics including residents' living conditions, ecological environment quality, lighting levels, urbanization degree, and accessibility. These features were aggregated at the DHS level and regressions were established with the IWI index to derive initial weights. Different regression models were further compared based on their R² values. As shown in Table 2, Partial Least Squares (PLS) regression demonstrated relatively stronger overall performance, thus serving as the baseline model for generating high-resolution wealth proxies. However, in practical implementation, to better align with diverse urban scenarios, we propose the domain-knowledge-adjusted PLS-A model. By modifying feature weights to better reflect the fundamental conditions of urban development levels, the PLS-A achieves improved spatial interpretability and more reasonable distribution patterns. Full implementation details and comparative results are provided in the Supplementary Note 3 and Figure S2. Table 2. Feature weight combinations and evaluation metrics across different methods. BA: Building Area; BD: Building Density; NDVI: Normalized Difference Vegetation Index; LIGHT: SDGSAT-1 Nighttime Lights; ROAD: OSM Road Density. All features are aggregated at the village level before performing regression analysis. Method Features Weights R 2 Explaining BA BD NDVI LIGHT ROAD POI X-Variance PCA 0.440 0.288 -0.490 0.407 0.462 0.323 \ 0.39 PLS 0.505 0.471 -0.237 0.331 0.462 0.381 0.56 0.60 Linear 0.250 0.127 0.031 0.041 0.102 0.076 0.51 \ Ridge 0.245 0.128 0.030 0.042 0.102 0.076 0.60 \ Lasso 0.226 0.075 0.000 0.000 0.088 0.040 0.56 \ PLS-A 0.505 0.235 0.237 0.165 0.231 0.191 \ \ Spatial Wealth Distribution and Inequality Analysis in Africa This analysis utilizes the African settlements dataset released by the Humanitarian OpenStreetMap Team. The original dataset contained 284,968 records, from which entries labeled "place=village or suburb or town" were selected. Points within a 2-kilometer radius were spatially aggregated using the DBSCAN clustering algorithm to ensure corresponding imagery did not overlap. Remote sensing imagery centered on each settlement was acquired using Google Earth Engine (GEE), ultimately yielding 59,998 valid image samples. Wealth estimates were generated for each village using the trained model. Within a 2.5 km radius around each settlement, the mean wealth and Gini coefficient were computed to assess local wealth levels and inequality. Due to variations in sampling and spatial coverage of OSM settlements—such as denser populations, higher infrastructure levels, and more frequent OSM annotations in West Africa—the 59,998 settlements were divided into four geographic regions: North Africa (north of the Sahara), West Africa, East Africa, and Southern Africa. This classification enables comparative analysis of spatial wealth and inequality across regions. West Africa accounts for the largest proportion of settlements (54%), followed by East Africa (31%), Southern Africa (8%), and North Africa (7%). The regional division is informed by geographic, institutional, and socioeconomic factors, with detailed criteria provided in Supplementary Table S6. Calculating Population Distribution by Wealth Tiers The wealth map was resampled to 100-meter resolution, and we combined the map with WorldPop (100m) [26] and the 2019 World Settlement Footprint (WSF2019) dataset to estimate the population size living within different wealth quantiles. WorldPop only counts residential regions filtered by the WSF2019 mask. The 20th and 80th percentiles were calculated at the national level. Based on these values, boundaries were established to delineate relatively underdeveloped regions (≤20%) and developed regions (≥80%), with the remaining population categorized as ordinary regions (20%–80%). To ensure consistency with official statistics, we further multiplied the proportion by the total national population as reported by the World Bank in 2020, thereby deriving the final stratified population estimate. Declarations Data availability The data product has been deployed on GEE and is publicly accessible at https://code.earthengine.google.com/ee8b591dc66c88a7158e3df82db911ea Code availability Code to replicate all findings in the paper are available at https://github.com/UsersLab-tx/Africa_Wealth_Mapping References Alkire, Sabina, and Maria Emma Santos. 2014. “Measuring Acute Poverty in the Developing World: Robustness and Scope of the Multidimensional Poverty Index.” World Development 59:251–274. https://doi.org/10.1016/j.worlddev.2014.01.026. Liu, Qi, Lei Gao, Zhaoxia Guo, Yucheng Dong, Enayat A. Moallemi, Sibel Eker, Jing Yang, Michael Obersteiner, and Brett A. Bryan. 2023. “Robust Strategies to End Global Poverty and Reduce Environmental Pressures.” One Earth 6 (4): 392–408. https://doi.org/10.1016/j.oneear.2023.03.007. Thomson, D.R.; Kuffer, M.; Boo, G.; Hati, B.; Grippa, T.; Elsey, H.; Linard, C.; Mahabir, R.; Kyobutungi, C.; Maviti, J.; et al. Need for an Integrated Deprived Area “Slum” Mapping System (IDEAMAPS) in Low- and Middle-Income Countries (LMICs). Soc. Sci. 2020, 9, 80. https://doi.org/10.3390/socsci9050080 Kuffer, M., Wang, J., Nagenborg, M., & Pfeffer, K. (2018). The scope of earth-observation to improve the consistency of the SDG slum indicator. ISPRS International Journal of Geo-Information, 7(11), 428. DOI:10.3390/ijgi7110428 Yeh, C., Perez, A., Driscoll, A. et al. Using publicly available satellite imagery and deep learning to understand economic well-being in Africa. Nat Commun 11, 2583 (2020). https://doi.org/10.1038/s41467-020-16185-w Tusting, L. S. et al. Mapping changes in housing in sub-Saharan Africa from 2000 to 2015. Nature 568, 391–394 (2019). Osgood-Zimmerman, A. et al. Mapping child growth failure in Africa between 2000 and 2015. Nature 555, 41 (2018). Frostad, J., Nguyen, Q., Baumann, M. M., Blacker, B., Marczak, L., Deshpande, A., Wiens, K., et al. Mapping development and health effects of cooking with solid fuels in low-income and middle-income countries, 2000–18: a geospatial modelling study. Lancet Glob. Health 10, e1522–e1535 (2022). DOI:10.1016/S2214-109X(22)00332-1 Graetz, N. et al. Mapping local variation in educational attainment across Africa. Nature 555, 48 (2018). Tatem AJ, Gething PW, Bhatt S, Weiss D and Pezzulo C (2013) Pilot high resolution poverty maps, University of Southampton/Oxford DOI: 10.5258/SOTON/WP00127 UN‐Habitat. (2004). The challenge of slums: global report on human settlements 2003. Management of Environmental Quality: An International Journal , 15 (3), 337-338. Sahn, D. E. & Stifel, D. Exploring alternative measures of welfare in the absence of expenditure data. Rev. Income Wealth 49 , 463–489 (2003). Filmer, D. & Pritchett, L. H. Estimating wealth effects without expenditure data-or tears: an application to educational enrollments in states of India. Demography 38 , 115–132 (2001). Deborah Johnston, Alexandre Abreu, The asset debates: How (not) to use asset indices to measure well-being and the middle class in Africa, African Affairs , Volume 115, Issue 460, July 2016, Pages 399–418, https://doi.org/10.1093/afraf/adw019 Sahn, D. E., & Stifel, D. C. (2000). Assets as a measure of household welfare in developing countries . Center for Social Development Research Report. Washington University in St. Louis.https://openscholarship.wustl.edu/cgi/viewcontent.cgi?article=1025&context=csd_research Kabudula, C. W., Houle, B., Collinson, M. A., & Kahn, K. (2017). Assessing changes in household socioeconomic status in rural South Africa, 2001–2013: A distributional analysis using household asset indicators. Social Indicators Research, 133(3), 1047–1073. https://link.springer.com/article/10.1007/s11205-016-1397-z Rutstein, S. O., & Staveteig, S. (2014). Making the Demographic and Health Surveys Wealth Index comparable . Resce, G. (2021). Wealth-adjusted Human Development Index . Journal of Cleaner Production . Alkire, S., Conconi, A., & Seth, S. (2014). Multidimensional Poverty Index 2014: Brief methodological note and results . Oxford Poverty and Human Development Initiative (OPHI). Smits, J., Steendijk, R. The International Wealth Index (IWI). Soc Indic Res 122, 65–85 (2015). https://doi.org/10.1007/s11205-014-0683-x Samper, J.; Shelby, J.A.; Behary, D. The Paradox of Informal Settlements Revealed in an ATLAS of Informality: Findings from Mapping Growth in the Most Common Yet Unmapped Forms of Urbanization. Sustainability 2020, 12, 9510. Lee, K., & Braithwaite, J. (2022). High-resolution poverty maps in Sub-Saharan Africa. World Development, 159, 106028. https://doi.org/10.1016/j.worlddev.2022.106028 Metzger, N., Daudt, R. C., Tuia, D. & Schindler, K. High-resolution population maps derived from Sentinel-1 and Sentinel-2. Remote Sens. Environ. (2024). Metzger, N., Vargas-Muñoz, J. E., Daudt, R. C., Kellenberger, B., Whelan, T. T. T., Ofli, F., ... & Tuia, D. (2022). Fine-grained population mapping from coarse census counts and open geodata. Scientific Reports, 12(1), 20085. Jacobs, N., Kraft, A., Rafique, M.U., Sharma, R.D., 2018. A weakly supervised approach for estimating spatial density functions from high-resolution satellite imagery. In: SIGSPATIAL International Conference on Advances in Geographic Information Systems. pp. 33–42. Tatem, A. J. (2017). WorldPop, open data for spatial demography. Scientific Data, 4, 170004. https://doi.org/10.1038/sdata.2017.4 Luo, E., Kuffer, M., & Wang, J. (2022). Urban poverty maps – From characterising deprivation using geo-spatial data to capturing deprivation from space . Sustainable Cities and Society, 84 , 104033. https://doi.org/10.1016/j.scs.2022.104033 Karsai, Márton et al. “A Comparative Analysis of Wealth Index Predictions in Africa between three Multi-Source Inference Models.” ArXiv abs/2408.01631 (2024): n. pag. Gorelick, N., Hancher, M., Dixon, M. et al. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment, 202, 18–27. https://doi.org/10.1016/j.rse.2017.06.031 Filippucci, P., Pulvirenti, L., Chini, M. et al. (2020). Sentinel-1 for mapping floods at the scale of the European continent. Remote Sensing, 12(2), 211. https://doi.org/10.3390/rs12020211 Microsoft. (2023). Global ML Building Footprints . GitHub. Retrieved April 15, 2025, from https://github.com/microsoft/GlobalMLBuildingFootprints Essential background in Pesaresi, M. et al. (2024) "Advances on the Global Human Settlement Layer by joint assessment of Earth Observation and population survey data", International Journal of Digital Earth, 17(1). Forget, Y., Shimoni, M., Gilbert, M., & Linard, C. (2021). Mapping 20 years of urban expansion in 45 urban areas of sub-Saharan Africa. Remote Sensing, 13(3), 525. https://www.mdpi.com/2072-4292/13/3/525 Ronneberger, Olaf et al. “U-Net: Convolutional Networks for Biomedical Image Segmentation.” ArXiv abs/1505.04597 (2015): n. pag. He, Kaiming et al. “Deep Residual Learning for Image Recognition.” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2015): 770-778. Abascal, Angela, et al. "“Domains of Deprivation Framework” for Mapping Slums, Informal Settlements, and Other Deprived Areas in LMICs to Improve Urban Planning and Policy: A Scoping Review." Computers, Environment and Urban Systems , vol. 93, 2022, p. 101770. https://doi.org/10.1016/j.compenvurbsys.2022.101770 Abascal, Angela, et al. "Identifying Degrees of Deprivation from Space Using Deep Learning and Morphological Spatial Analysis of Deprived Urban Areas." Computers, Environment and Urban Systems, vol. 95, 2022, p. 101820. https://doi.org/10.1016/j.compenvurbsys.2022.101820 Li, Chengxiu, et al. "Slum and Urban Deprivation in Compacted and Peri-Urban Neighborhoods in Sub-Saharan Africa." Sustainable Cities and Society , vol. 99, 2023, p. 104863. https://doi.org/10.1016/j.scs.2023.104863 Thomson DR, Kuffer M, Boo G, Hati B, Grippa T, Elsey H, Linard C, Mahabir R, Kyobutungi C, Maviti J, et al. Need for an Integrated Deprived Area “Slum” Mapping System (IDEAMAPS) in Low- and Middle-Income Countries (LMICs). Social Sciences . 2020; 9(5):80. https://doi.org/10.3390/socsci9050080 UNDP (United Nations Development Programme). (2020). Angola Human Development Report 2020. Home | United Nations Development Programme UN Economic Commission for Africa (ECA), The Economic Implications of the Oil Sector in Sub-Saharan Africa: Angola , 2015. Beegle, K., & Christiaensen, L. (2019). Accelerating Poverty Reduction in Africa. Washington, DC: World Bank. https://doi.org/10.1596/978-1-4648-1232-3 Li, C., Yu, L., Ndugwa, R. et al. Mapping urban slums and their inequality in sub-Saharan Africa. Nat Cities (2025). https://doi.org/10.1038/s44284-025-00276-0 Christiaensen, L., & Todo, Y. (2014). Poverty reduction during the rural–urban transformation – The role of the missing middle . World Development, 63, 43–58. Fosu, A. K. (2017). Growth, inequality, and poverty reduction in developing countries: Recent global evidence . Research in Economics, 71(2), 306–336. Banerjee, A. & Duflo, E. Poor Economics (2011). Classic evidence on poverty traps and the effectiveness of education/basic-services/SME interventions; your policy framing aligns with these micro-foundations. Alkire, S., & Santos, M. E. (2014). Measuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index . World Development, 59, 251–274. https://doi.org/10.1016/j.worlddev.2014.01.026 Satterthwaite, D. (2017). The impact of urban development on risk in sub-Saharan Africa’s cities with a focus on small and intermediate urban centres . International Journal of Disaster Risk Reduction, 26, 16–23. https://doi.org/10.1016/j.ijdrr.2017.09.025 Wang, L., Long, T., Jiang, W., Adam, E., Wen, C., Jiao, W., & He, G. (2025). Economic well-being assessment: a review of traditional and remote sensing approaches. International Journal of Digital Earth , 18 (1). https://doi.org/10.1080/17538947.2025.2504137 Espey, J. et al. Data for development: a needs assessment for SDG monitoring and statistical capacity development. Sustain. Dev. Solut. Netw. http://unsdsn.org/wp-content/uploads/2015/04/Data-for-Development-Full-Report.pdf (2015). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryMaterials.docx Supplementary Materials Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8234710","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":558284219,"identity":"b0ba1db2-9830-41b7-98f4-116be716d4df","order_by":0,"name":"Qingyan Meng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxUlEQVRIiWNgGAWjYPACGyBObCBJSxrpWg4DcQKRanXbew+/eNt2Pp+/PbmB4UcNg7w5IS1mZ86lWc5tu20548zDBsaeYwyGOwm5z+xGjpkxb9ttA4YbQL/wNjAkGBwgpOX+G5CWcwbyQC2Mf4nScoPH+DFv2wEDA6AWZuJsOZNjxjjnXLKBIdAvh2WOSRhuIKjl+BnjD2/K7Azkjqc/fPimxkaeoC1AwCbBywZhARVLEFYPBMwfeP4QpXAUjIJRMApGKgAABrxEoyE87cwAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-5440-4081","institution":"Aerospace Information Research Institute, Chinese Academy of Sciences","correspondingAuthor":true,"prefix":"","firstName":"Qingyan","middleName":"","lastName":"Meng","suffix":""},{"id":558284220,"identity":"b3cda124-990c-4484-bec3-dc8f8269cfa7","order_by":1,"name":"Xin Tan","email":"","orcid":"","institution":"Aerospace Information Research Institute, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Tan","suffix":""},{"id":558284221,"identity":"b74d15e5-3411-4e6a-8089-396654c6d6e7","order_by":2,"name":"Qikang Zhao","email":"","orcid":"https://orcid.org/0009-0005-5969-3723","institution":"Aerospace Information Research Institute, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Qikang","middleName":"","lastName":"Zhao","suffix":""},{"id":558284222,"identity":"3f00f8fe-3e8f-4943-b05f-6af0dba0978c","order_by":3,"name":"Linlin Zhang","email":"","orcid":"","institution":"Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing","correspondingAuthor":false,"prefix":"","firstName":"Linlin","middleName":"","lastName":"Zhang","suffix":""},{"id":558284223,"identity":"ddcce0e8-bcc6-4b96-8973-1762520cd244","order_by":4,"name":"Liqiang Zhang","email":"","orcid":"https://orcid.org/0000-0002-4175-7590","institution":"State Key Laboratory of Remote Sensing and Digital Earth, Faculty of Geographical Science, Beijing Normal University, Beijing","correspondingAuthor":false,"prefix":"","firstName":"Liqiang","middleName":"","lastName":"Zhang","suffix":""},{"id":558284224,"identity":"3a8aa606-fa9f-4b40-883f-1b0bac53998b","order_by":5,"name":"Chuanglin Fang","email":"","orcid":"https://orcid.org/0000-0001-6043-8491","institution":"Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Chuanglin","middleName":"","lastName":"Fang","suffix":""}],"badges":[],"createdAt":"2025-11-29 06:10:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8234710/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8234710/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97952363,"identity":"81ce6e19-370a-453c-a17c-382b9ad2c7c5","added_by":"auto","created_at":"2025-12-11 07:23:02","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8413470,"visible":true,"origin":"","legend":"","description":"","filename":"HighresolutionwealthmapsrevealAfricaXXXsnonlineartrajectoryofdevelopmentandinequality1122.docx","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/6265425b070e9e53e3872fc9.docx"},{"id":97952347,"identity":"ad0eb964-0870-49a8-9fa9-c0c157100e96","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7275,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2595450.json","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/2a3282132b1cdc8389ee0f49.json"},{"id":97952355,"identity":"12305186-64b6-4a97-bd2e-3e794a4108ab","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11915945,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/f3236bc39f46c5dfed899266.docx"},{"id":97952346,"identity":"61cbf490-a0ab-4bb4-9fa0-3806e88b41f2","added_by":"auto","created_at":"2025-12-11 07:23:00","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115404,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25954500enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/c2ed4ad4684ae277d6772ffd.xml"},{"id":98422808,"identity":"d461e818-83b5-49f0-929b-63867fa17d5e","added_by":"auto","created_at":"2025-12-17 16:31:32","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":542888,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/1fc19eca915b60c34a5487a5.png"},{"id":98423668,"identity":"e2f16cfe-081d-4dd7-88ef-946b030fdc48","added_by":"auto","created_at":"2025-12-17 16:32:30","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":953227,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/ee1402f3b1c04ba30c6ad1d8.png"},{"id":97952364,"identity":"5e8a53b5-5902-4b1f-9b55-67fc70a65a4e","added_by":"auto","created_at":"2025-12-11 07:23:03","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2170752,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/90911c317cbf26228ba50828.png"},{"id":98422410,"identity":"3f33219d-87cb-4a6e-bada-995b68fa4cb8","added_by":"auto","created_at":"2025-12-17 16:31:00","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":181693,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/a290db44bec1b739737bfa96.png"},{"id":97952367,"identity":"b7afb6e5-13b0-46bd-b842-55ff33863e32","added_by":"auto","created_at":"2025-12-11 07:23:03","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3590717,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/58867decd62cdd5ce130677b.png"},{"id":98422449,"identity":"492dd166-1534-4fb4-9120-4c0f8303d016","added_by":"auto","created_at":"2025-12-17 16:31:03","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":843799,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/f88ee6b1b91173b3b372eaad.png"},{"id":97952352,"identity":"d589b6a7-1a96-4116-92d5-fad7a92bc10d","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111277,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/611a9336c7b1054be5994a47.png"},{"id":97952356,"identity":"31cded47-4989-4aa8-8d16-8c9f93e8fffe","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":157409,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/90133b47b045c548cf9eaef3.png"},{"id":97952351,"identity":"e3e85fcc-eb65-47a2-86d1-78108103135a","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342637,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/012fbbec815575f1481fb667.png"},{"id":98423470,"identity":"b4280803-8c68-41e1-9752-2b72a94925be","added_by":"auto","created_at":"2025-12-17 16:32:16","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43443,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/c7d5f2d315e1a656fda3d202.png"},{"id":97952365,"identity":"203e3714-58ce-4944-9fe5-378d00b3d87c","added_by":"auto","created_at":"2025-12-11 07:23:03","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":714306,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/ad1662d123cdc8022e0c5b58.png"},{"id":98422042,"identity":"8e162a73-f2a7-4baa-88df-a8a54a199f02","added_by":"auto","created_at":"2025-12-17 16:30:21","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":156400,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/551234d5ab9e3f7b7517d2f4.png"},{"id":97952360,"identity":"54dab1dd-d7e6-470c-967a-5db5c69384c9","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":113740,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25954500structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/38c5d1ec60949224ecf753fd.xml"},{"id":97952359,"identity":"cbce2c73-81ff-45c3-9ce3-6ea815136c8b","added_by":"auto","created_at":"2025-12-11 07:23:01","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":127112,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/c23768b1c2641c0904791696.html"},{"id":98422544,"identity":"00461105-acac-421d-9a23-6de32252f902","added_by":"auto","created_at":"2025-12-17 16:31:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":382081,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAccuracy Validation.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003e Cross-validation accuracy (Mean Absolute Error (MAE) by country, urban vs. rural); \u003cstrong\u003e(b)\u003c/strong\u003e National error map (R\u003csup\u003e2\u003c/sup\u003e and MAE); \u003cstrong\u003e(c)\u003c/strong\u003e Validation of predicted IWI with DHS-derived IWI at village level; \u003cstrong\u003e(d)\u003c/strong\u003e and \u003cstrong\u003e(e)\u003c/strong\u003e: Comparison of our predicted results with Lee et al. [22] and Yeh et al. [5] at the village level.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/517b90ff6847977fbf4095c1.png"},{"id":97952344,"identity":"01687394-1428-485e-825a-48e2e766e2df","added_by":"auto","created_at":"2025-12-11 07:23:00","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":579895,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpatial Wealth Index and Gini Coefficient across Africa’s populated settlements. (a)\u003c/strong\u003eKDE curves of the Spatial Wealth Index (SWI) by region; \u003cstrong\u003e(b)\u003c/strong\u003e Scatter plot of mean wealth vs. Gini coefficient; \u003cstrong\u003e(c)\u003c/strong\u003e SWI map aggregated at the 2 km settlement level with regional stacked-bar charts, showing the ratio of settlements across different wealth tiers; \u003cstrong\u003e(d)\u003c/strong\u003e Gini coefficient map with regional boxplots; \u003cstrong\u003e(e)\u003c/strong\u003e Per-country wealth distributions with Gini coefficient markers.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/f22e5932dff9b6c62d73775f.png"},{"id":98423815,"identity":"726aed24-354a-434d-ae59-0a5785b5e8dd","added_by":"auto","created_at":"2025-12-17 16:32:38","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1100777,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpace Wealth Index map—Angola's performance at 2020\u003c/strong\u003e. The side curves represent the aggregated index values along latitude and longitude. Population curves are derived from WorldPop [26].\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/0db61d170ee2567ca4cae790.png"},{"id":98422805,"identity":"bb473554-c06d-4621-9c23-f6c9161cefc6","added_by":"auto","created_at":"2025-12-17 16:31:32","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":128898,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePopulation Distribution by Wealth Tiers cross 30 African Countries in 2020.\u003c/strong\u003e Using the 20th and 80th percentiles of each country’s wealth map, we define relatively underdeveloped (≤20%), developed (≥80%), and intermediate (20–80%) regions. WorldPop data are then used to estimate the proportion of populations residing in each region.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/0137e3b8786e77459a4a69a3.png"},{"id":97952366,"identity":"7cfdf64b-c287-4273-8d9d-7f90764900b2","added_by":"auto","created_at":"2025-12-11 07:23:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1579966,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHigh-resolution Spatial Wealth Index Maps across African cities. (Left)\u003c/strong\u003e The 10 m Spatial Wealth Index map of Luanda, Angola, with insets \u003cstrong\u003e(a–d)\u003c/strong\u003e highlighting distinct urban scenes:\u003cstrong\u003e(a)\u003c/strong\u003e the central business district, \u003cstrong\u003e(b)\u003c/strong\u003e a medium-density residential–commercial mixed district, \u003cstrong\u003e(c)\u003c/strong\u003e a medium-density residential neighborhood, \u003cstrong\u003e(d)\u003c/strong\u003e and a low-income informal settlement. \u003cstrong\u003e(Right and bottom panels)\u003c/strong\u003e Comparison of the predicted wealth index \u003cstrong\u003e(first column)\u003c/strong\u003e with \u003cstrong\u003e(second column) \u003c/strong\u003ethe 100 m Spatial Wealth Index product from Li et al. [43], and \u003cstrong\u003e(third column) \u003c/strong\u003enighttime light intensity across representative urban scenes in Luanda, as well as additional examples from Ghana and Kenya. The figure illustrates how high-resolution wealth mapping captures fine-scale socioeconomic variation beyond what can be inferred from traditional proxies.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/e12876980f2102457aefec58.png"},{"id":98423771,"identity":"cae37e56-08e7-467d-b773-c4c967b3debd","added_by":"auto","created_at":"2025-12-17 16:32:35","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":511947,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTechnical framework. \u003c/strong\u003eA: constructing household asset wealth index, B: constructing spatial wealth index proxy, C: multi-source remote sensing imagery to Unet, D: loss supervision strategy.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/527b3d9c638117815deacf64.png"},{"id":98622388,"identity":"ffa4ed9e-fa29-4101-bc59-8d4d7adb5ba3","added_by":"auto","created_at":"2025-12-19 16:53:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5139509,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/7e4564c1-d146-4e1a-b8be-731ae5a5acd5.pdf"},{"id":97952362,"identity":"3ad948b0-278c-419c-9f18-308dd555eeba","added_by":"auto","created_at":"2025-12-11 07:23:02","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11915945,"visible":true,"origin":"","legend":"Supplementary Materials","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-8234710/v1/9a0b9396b109f85b80db13af.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"High-resolution wealth maps reveal Africa’s nonlinear trajectory of development and inequality","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eSocial and economic well-being is a fundamental reference for understanding development and guiding policies. Globally, poverty and inequality are tightly intertwined: economic growth has not provided equal opportunities, instead, in many countries, the urban-rural divide and regional disparities have been further widened. Underdeveloped regions face not only poverty but also multiple deficiencies in infrastructure, services, and basic living conditions [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Such multidimensional deprivation often overlaps with inequality, creating a vicious cycle of poverty and social exclusion. Recent studies [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e] suggest that, even under optimistic scenarios, the global goal of poverty eradication will be delayed by more than two decades, underscoring the urgency for intensified efforts. However, there is still little clarity on how many people live in underdeveloped regions and how poverty and inequality interact[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. This knowledge gap limits accurate assessments of development conditions, weakens policy design, and hinders the achievement of the Sustainable Development Goals (SDGs)[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe Demographic and Health Surveys (DHS) Program has conducted more than 400 surveys across over 90 countries, providing accurate and representative data on population, health, HIV, and nutrition (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.dhsprogram.com/\u003c/span\u003e\u003cspan address=\"https://www.dhsprogram.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). While these surveys help fill critical data gaps, their infrequent updates\u0026mdash;often three years apart within the same country\u0026mdash;limit their ability to capture dynamic socioeconomic conditions across Africa [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. To address this limitation, recent studies have sought to integrate DHS data with Satellite-observation methods characterized by rapid updates and broad coverage, producing large-scale socioeconomic maps such as household asset wealth [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], access to improved housing [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], child development [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], the impacts of solid cooking fuel use [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], educational attainment [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e], and poverty rates [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. These efforts represent important advances toward spatializing well-being, yet most are still restricted to the village scale (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), limiting their ability to reveal intra-community disparities or to support localized decision-making.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAn Overview of current map products related to poverty or economic well-being.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMap Product\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCoverage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDuration\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLimitation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlobal Subnational Atlas of Poverty (GSAP)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSubnational poverty estimates\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGlobal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2010, 2019, 2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eWorldBank\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLack of ability to analyze further\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eImproved Housing Map\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eProportion of households living in improved housing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSub-Saharan Africa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2000, 2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTusting et al.[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5 km gridded data; limited applicability in community-level planning\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInformal Atlas (AoI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVector maps of informal settlements\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e147 cities, 102 countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eselected years from 2000 to 2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSamper et al.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCrowdsourced data; partial representation of informal areas\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePilot high resolution poverty maps\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGridded estimates of poverty population\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTatem et al. [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEarly-stage poverty mapping initiative\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh-resolution poverty maps in Sub-Saharan Africa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAsset-based wealth index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e44 Sub-Saharan African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLee et al. [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAt the settlement level\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMapping urban slums and their inequality in sub-Saharan Africa\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e100m Slum Mapping and Wealth Index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32 Sub-Saharan African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLi et al. [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eOnly in urban areas\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eMost low- and middle-income countries (LMICs) are undergoing or will soon experience rapid urbanization, accompanied by the unexpected expansion of slum-like communities[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. To support the transformation of these areas into \"inclusive, safe, resilient, and sustainable\" communities according to the SDG11, understanding residents' economic and living conditions is an urgent priority. Household Asset Index (HAI) widely recognized as one of the core indicators for measuring well-being, reflecting an individual's satisfaction with their material conditions, encompassing all aspects of income, assets, access to basic services[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e][\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. The index is specifically constructed based on possession of physical assets (e.g., housing, vehicles, durable goods), housing conditions (e.g., water supply, electricity, sanitation), and essential utilities. Due to its independence from direct monetary income data, the HAI is particularly effective in regions where income is difficult to measure or household-level data are scarce. These advantages have contributed to its widespread adoption in socioeconomic surveys across many African countries[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e][\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e][\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Internationally, multiple methods have been proposed to measure the HAI, including DHS Wealth Index [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], Human Development Index [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], Multidimensional Poverty Index [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], and International Wealth Index (IWI) [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Among them, the IWI stands out for adjusting cross-country differences in survey design and asset variables, allowing for comparable assessments at both regional and global levels. Recent studies[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] further demonstrate that IWI, when integrated with Earth observation and micro-survey data, provides a more accurate representation of household wealth distribution, especially in Africa. Building on this evidence, we adopt IWI as the target indicator for high-resolution mapping.\u003c/p\u003e \u003cp\u003eTo gain deeper insights into Africa\u0026rsquo;s current micro-level socioeconomic conditions, we extend indicators traditionally constrained to the settlement scale onto high-resolution maps. Specifically, drawing on approaches from fine-scale population mapping researches [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e][\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e][\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], we develop a deep learning framework that combines regional survey indicators with fine-grained spatial proxies: the village-level (~\u0026thinsp;5km) IWI scores, constructed from DHS data, are projected onto high-resolution (10-m) maps under the guidance of spatial wealth proxies. These proxies are composite indicators constructed from 10-m features of housing quality, ecological environment, and socioeconomic activity. Applying this framework, we produced the first 10-m wealth maps for 30 African countries, providing unprecedented fine-grained insights into population distributions across different levels of development. By aggregating these fine-grained development conditions, analysis of more than 60,000 settlements further reveals marked regional disparities and a nonlinear U-shaped pattern between development and inequality. These patterns not only help distinguish divergent national development pathways but also underscore the compounded risks of poverty and social exclusion under extreme deprivation. Overall, our findings uncover hidden dynamics of development and inequality in Africa and provide a new foundation for poverty alleviation and equitable infrastructure planning.\u003c/p\u003e"},{"header":"2. Result","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Performance and Accuracy Analysis\u003c/h2\u003e \u003cp\u003eTo validate the accuracy of the proposed model, we conducted a multi-scale evaluation across 30 countries in sub-Saharan Africa. The assessment metrics are based on the average results of five-fold cross-validation. The model demonstrates strong predictive performance for the village-level IWI derived from DHS data (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec), achieving an overall coefficient of determination (R\u0026sup2;) of 0.74 and a mean absolute error (MAE) of 0.24, indicating the effectiveness in estimating wealth across countries. Overall, we find that predictions are more stable in rural areas than in urban samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). In several countries, such as Burundi (BDI), Zambia (ZMB), and Malawi (MWI), due to the complex spatial heterogeneity of urban environments, urban prediction errors are noticeably higher. Spatially (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb), accuracy is significantly higher in West Africa compared to East and Southern Africa with broad national coverage, including six countries achieved an R\u0026sup2; above 0.8, and two-thirds of the countries exceeded an R\u0026sup2; of 0.7. In contrast, South Africa (ZAF) recorded the lowest accuracy, with an R\u0026sup2; of 0.47. We further validated our method by comparing it against leading approaches in spatial wealth prediction. The results show high consistency with the cross-country models proposed by Lee et al. [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] and Yeh et al. [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], both yielding R\u0026sup2; values above 0.75.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Wealth Distribution and Inequality in Africa\u003c/h2\u003e \u003cp\u003eTo reveal Africa\u0026rsquo;s overall socioeconomic conditions, we mapped the distribution of development level (mean wealth) and internal inequality (Gini coefficient) across nearly 60,000 densely populated settlements, each representing a two-kilometer situation. The results show a significant association between development level and internal inequality (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, r\u0026thinsp;=\u0026thinsp;0.59), coupled with a characteristic nonlinear U-shaped curve. Notably, the segmented regression identified two critical inflection points (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, c\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e = \u0026minus;\u0026thinsp;0.66, c2 = \u0026minus;\u0026thinsp;0.38), with about 63% of settlements concentrated at the \u0026ldquo;bottom of the ladle\u0026rdquo; (c1\u0026thinsp;\u0026lt;\u0026thinsp;wealth\u0026thinsp;\u0026lt;\u0026thinsp;c2), characterized by low wealth and low inequality\u0026mdash;a state of \u0026ldquo;equilibrated poverty\u0026rdquo;. Beyond this range, settlements diverge along two trajectories: one follows a \u0026ldquo;high-wealth\u0026ndash;high-inequality\u0026rdquo; pathway (wealth\u0026thinsp;\u0026gt;\u0026thinsp;c2), where urbanization drives rapid growth alongside widening disparities; the other falls into a \u0026ldquo;low-wealth\u0026ndash;high-inequality\u0026rdquo; state (wealth\u0026thinsp;\u0026lt;\u0026thinsp;c1), marked by the simultaneous intensification of poverty and inequality.\u003c/p\u003e \u003cp\u003eOur results also reveal evident regional disparities (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). Specifically, North Africa accounts for only 7% of settlements. Given their concentrated distribution along the coastal economic belt, it exhibits the highest average wealth and inequality, with 43% of settlements falling within the top 20% of the Africa wealth distribution. In contrast, West Africa contains more than half of all settlements (54%). Due to its widespread rural settlements, it exhibits the lowest average wealth and about 27% of settlements in the bottom 20%, reflecting widespread poverty. At the national scale (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee), quintile-based analysis further identifies countries exhibiting two distinct development patterns (See Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). \u0026ldquo;High-wealth\u0026ndash;high-inequality\u0026rdquo; countries such as Gabon and Rwanda reflect a concentrated development pattern, whereas \u0026ldquo;low-wealth\u0026ndash;high-inequality\u0026rdquo; countries such as Togo and Gambia exhibit compounded deprivation risks.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Space Wealth Index map - Angola\u003c/h2\u003e \u003cp\u003eWe present a Spatial Wealth Index map, Angola as an example, to further illustrate development disparities at the national level (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Angola's development pattern exhibits a \u0026ldquo;coastal-to-interior\u0026rdquo; gradient, with the aggregation curve along the latitude-longitude axis indicating a significant synergistic distribution relationship between wealth and population (r\u0026thinsp;=\u0026thinsp;0.81, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). In the capital Luanda, one-quarter of the population accounts for nearly 60% of the nation's wealth index accumulation (calculation method see Supplementary Note 4), illustrating Angola's highly concentrated development pattern. Central provinces, such as Malanje and Cuando Cubango, exhibit generally moderate levels of development, with wealth concentrated in their core areas. Eastern provinces like Moxico and Alto Zambeze, however, remain in a state of widespread poverty, with reported poverty rates exceeding 50% and reaching over 70% in some regions [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Spatial autocorrelation analysis further indicates that poverty settlements (bottom 20% in Angola\u0026rsquo;s wealth distribution) are significantly clustered in rural regions of Huambo, Malanje, and Lunda Norte (Moran\u0026rsquo;s I\u0026thinsp;\u0026asymp;\u0026thinsp;0.41, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), where structural constraints\u0026mdash;including inadequate infrastructure, sparse population, and complex terrain, which hinder development opportunities[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Population Distribution by Wealth Tiers cross 30 African Countries\u003c/h2\u003e \u003cp\u003eBeyond the Angola case, we further analyzed the proportion of the population across different wealth tiers in 30 African countries (see Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e for details). The total population of these countries is approximately 940\u0026nbsp;million, with 36% (about 338\u0026nbsp;million) residing in the least developed regions (the bottom 20% in wealth distribution). This proportion is consistent with estimates from the World Bank Poverty and Inequality Platform (PIP), which report that around 35% of Sub-Saharan Africa\u0026rsquo;s population lived below the \u003cspan\u003e$\u003c/span\u003e1.90/day extreme poverty line in 2019 [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. This consistency aims to validate the effectiveness of using the lowest wealth quintile as a proxy for poverty and does not imply equivalence to the extreme poverty line. At the national level, countries such as Gambia (72.5%), Senegal (68.7%), Sierra Leone (66.8%), and Togo (64.7%) exhibit poverty rates exceeding 60%, far above the regional average, underscoring the acute vulnerability of West Africa and parts of Southern Africa and the urgent priority for targeted policy interventions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eAlthough settlement-level maps can reflect basic socioeconomic conditions [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e][\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], high-resolution wealth maps provide additional insights into the internal disparities across different functional zones (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea\u0026ndash;d), consistent with objective socioeconomic realities. Notably, informal settlements exhibit substantially lower wealth, demonstrating the map\u0026rsquo;s ability to detect spatial deprivation in urban environments (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed). Compared to the 100-m wealth maps restricted to major urban areas by Li et al. [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e], our 10-m maps not only provide finer spatial details but also achieve full urban\u0026ndash;rural coverage. This makes them an effective and low-cost complement to traditional household surveys, which typically require investments of nearly one billion USD annually across low-income countries to monitor indicators related to the Sustainable Development Goals (SDGs)[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Moreover, wealth maps can be viewed as a comprehensive reflection of socioeconomic development. When combined with population data, they can answer specific questions such as \u0026ldquo;Where does poverty exist and how massive is it\u0026rdquo;\u0026mdash; spatial and structural patterns invisible in coarse-resolution data [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e][\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Importantly, our framework remains robust even under limited data conditions: when trained with only 10% of the samples, the model still performs well (Table S3, test R\u0026sup2; = 0.54). Beyond wealth estimation, this framework can be extended to other socioeconomic indicators, including the Multidimensional Poverty Index (MPI), Asset Index Score, household consumption expenditure, and per capita consumption.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePoverty in Africa remains extensive, with many countries exhibiting an imbalanced population distribution of \u0026ldquo;large proportions at both ends and a hollowed middle,\u0026rdquo; reflecting the dual challenges of current development. By aggregating these fine-grained differences, we found that settlement-level trajectories are more complex than commonly assumed: although economic development tends to widen micro-level disparities, the process is not linear\u0026mdash;neither \u0026ldquo;the richer, the more unequal\u0026rdquo; nor \u0026ldquo;the poorer, the more equal.\u0026rdquo; Instead, it follows a nonlinear U-shaped curve [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e], emphasizing the dominance of a large \u0026ldquo;equilibrated poverty\u0026rdquo; group that consistent with previously mentioned conditions of poverty and inequality, while offering new insights into their underlying dynamics. For this \u0026ldquo;equilibrated poverty\u0026rdquo; group, policy priorities should focus on enhancing inclusive efficiency, expanding education, basic services, and small and medium enterprise (SME) support\u0026mdash;to help populations escape the \u0026ldquo;poverty trap\u0026rdquo; and move into middle-wealth levels [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. For those still in extreme poverty, the priority lies in improving infrastructure and living conditions to mitigate multidimensional deprivation and avoid compounded cycles of poverty and inequality [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e][\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. Meanwhile, in rapidly urbanizing communities, wealth accumulation is evident but internal inequality is also intensifying, calling for equitable infrastructure distribution and institutional redistribution to prevent excessive disparities and the growth of informal settlements [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eFinally, it is expected to offer a new instrument for aligning macro-level policy goals with micro-level realities. Specifically, they enable cross-validation between national targets (e.g., ~\u0026thinsp;60% located in the \u0026ldquo;ladle bottom\u0026rdquo;) and micro-level patterns (e.g., the bottom 20% population share), ensuring that poverty alleviation and equity strategies are both evidence-based and geographically precise[\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e].\u003c/p\u003e"},{"header":"4. Methods","content":"\u003cp\u003eThis study proposes a high-resolution spatial wealth estimation framework that integrates multi-source remote sensing imagery with auxiliary data. Overall, we design a deep learning approach jointly supervised by village-level wealth data (5 km \u0026times; 5 km) and a spatial wealth proxy (10 m \u0026times; 10 m). The technical framework comprises four main components:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA. International Wealth Index (IWI):\u003c/strong\u003e We construct a standardized wealth index using principal component analysis (PCA) based on asset-related variables from Demographic and Health Surveys (DHS), including five types of durable goods, two public services, and three housing characteristics. This index serves as the supervisory signal for model training and evaluation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB. Spatial Wealth Proxy Features:\u0026nbsp;\u003c/strong\u003eHigh-resolution proxy features\u0026mdash;including built-up area, building density, NDVI, nighttime lights, road network density, and POI density\u0026mdash;are used to construct a \u0026quot;spatial wealth proxy map,\u0026quot; providing pixel-level spatial guidance during model training.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eC. Multi-Source Input Imagery:\u003c/strong\u003e The model integrates Sentinel-1 radar imagery, Sentinel-2 multispectral data, and SDGSAT-1 nighttime light imagery to enrich spatial context and semantic features, which are fed into a UNet-based architecture.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eD. Loss Monitoring Strategy:\u003c/strong\u003e A joint supervision mechanism is established, combining region-level wealth index and the spatial distribution characteristics of the proxy map to guide model optimization. This strategy enhances both predictive accuracy and spatial coherence of the output.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDHS Survey Data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe collected 51 survey rounds conducted between 2015 and 2024 across 30 African countries. Each country is typically surveyed every 3 to 5 years. For privacy protection, household-level data are aggregated at the cluster level, usually corresponding to a village and each cluster is geo-referenced by GPS coordinates. After preprocessing, approximately 23,700 valid clustered datasets were ultimately formed. It should be noted that to protect respondent privacy, the DHS offsets cluster coordinates by up to 10 kilometers in rural areas and 2 kilometers in urban areas. Although this spatial displacement introduces uncertainty into the mapping result, previous studies [5] have shown that the wealth index remains robust under such displacement, with R\u0026sup2; variations less than 0.07 when coordinates are shifted by 2.5 km.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWealth Index Construction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing prior studies [22], we calculated the IWI for each household based on ownership of ten specific assets: five durable goods (television, refrigerator, telephone, bicycle, and car), access to two public services (water and electricity), and three housing characteristics (number of bedrooms, flooring material, and quality of sanitation facilities). The index was constructed using the first principal component derived from a principal component analysis (PCA). The IWI accounts for differences in survey design and asset variables across nations, enabling the construction of globally or regionally consistent wealth maps.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSatellite Imagery\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGoogle Earth Engine [29] provides archival surface reflectance imagery from Sentinel-2 and Sentinel-1 satellites since 2015. For Sentinel-2, we selected the RGBN bands. Sentinel-1 data were extracted from the Interferometric Wide (IW) mode, including both VV and VH dual-polarization channels. To enhance data quality, we removed pixels with low signal-to-noise ratios (below \u0026ndash;30 dB), then computed multi-channel mean composites and applied median filtering for noise reduction [30].\u003c/p\u003e\n\u003cp\u003eFor each DHS cluster, we centered a 5120-meter cropping window on the provided GPS location to cover DHS spatial displacement (typically less than 2 km). This is a widely used preprocessing strategy, although it may introduce some noise, we consider local conditions to be generally stable. Notably, the collected imagery (excluding the nighttime light imagery) was gathered within one year before and after the survey period. It was subsequently processed using median filtering to ensure the highest-quality imagery closest to the survey timeframe.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuxiliary Data and Dataset Partitioning\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBuilding footprint data were obtained from Microsoft\u0026rsquo;s global building footprint product [31], this dataset effectively captures building form and density characteristics across Africa. Additional auxiliary data include road networks and points of interest (POIs) from the Humanitarian OpenStreetMap Team (HOT); Normalized Difference Vegetation Index (NDVI) derived from Sentinel-2 imagery; building area grids from the Global Human Settlement Layer (GHSL) [32]; and 10-meter nighttime light imagery from the SDGSAT-1 satellite, captured in 2021. Although these data were collected at different times, the built environment exhibits high structural stability over the short term (3\u0026ndash;5 years) [33] and allows for acceptable tolerance to temporal mismatch in features such as buildings and infrastructure.\u003c/p\u003e\n\u003cp\u003eDuring dataset construction, Sentinel imagery and auxiliary data were first paired at the sample level. Samples with severe cloud contamination were removed, resulting in a final dataset of 21,233 matched samples. This dataset was partitioned using a five-fold cross-validation strategy. First, five test sets were created, each derived from data collected in at least three independent national surveys and containing 1,800 to 2,300 samples (approximately 10% of the total sample size). None of these test sets were used for training. The remaining data were randomly split into training and validation sets in an 8:2 ratio. Detailed partitioning information is provided in Table S3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRegion-Level and Pixel-Level Guided Deep Learning Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe propose a CNN architecture that integrates multi-source remote sensing data. The network is built upon the UNet backbone [34], with a ResNet-50 encoder [35] for feature extraction from Sentinel-1 and Sentinel-2 imagery. Additionally, we design a feature enhancement module, LightNet, to encode SDGSAT-1 data, which is subsequently fused with the main feature stream. The architecture is described as Supplementary Note 1 and Note 2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpatial Wealth Proxy Construction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe IDEAMAPS framework [36] constructs a multidimensional conceptual system for urban poverty, encompassing socioeconomic conditions, infrastructure and services, environmental quality, and unplanned urbanization. While it has demonstrated effectiveness in localized areas such as Nairobi [27], its application at larger scales remains constrained by limited data availability and uneven label distribution. To address this, we focus on extracting indicators related to urban form and socioeconomic status [37][38][39], supporting large-scale spatial modeling across Africa.\u003c/p\u003e\n\u003cp\u003eSpecifically, the spatial wealth proxy index was constructed using six high-resolution features\u0026mdash;building area, building density, vegetation index, nighttime light intensity, road network density, and point of interest density\u0026mdash;to reflect multidimensional characteristics including residents\u0026apos; living conditions, ecological environment quality, lighting levels, urbanization degree, and accessibility. These features were aggregated at the DHS level and regressions were established with the IWI index to derive initial weights. Different regression models were further compared based on their R\u0026sup2; values.\u003c/p\u003e\n\u003cp\u003eAs shown in Table 2, Partial Least Squares (PLS) regression demonstrated relatively stronger overall performance, thus serving as the baseline model for generating high-resolution wealth proxies. However, in practical implementation, to better align with diverse urban scenarios, we propose the domain-knowledge-adjusted PLS-A model. By modifying feature weights to better reflect the fundamental conditions of urban development levels, the PLS-A achieves improved spatial interpretability and more reasonable distribution patterns. Full implementation details and comparative results are provided in the Supplementary Note 3 and Figure S2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Feature weight combinations and evaluation metrics across different methods.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBA: Building Area; BD: Building Density; NDVI: Normalized Difference Vegetation Index; LIGHT: SDGSAT-1 Nighttime Lights; ROAD: OSM Road Density. All features are aggregated at the village level before performing regression analysis.\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eMethod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"6\"\u003e\n \u003cp\u003eFeatures Weights\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eExplaining\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNDVI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLIGHT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROAD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePOI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eX-Variance\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePCA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.440\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.288\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.490\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.407\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.462\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.323\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePLS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.505\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.471\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.237\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.331\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.462\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.381\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.60\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLinear\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.250\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.127\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.031\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.102\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.076\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eRidge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.245\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.128\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.030\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.102\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.076\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLasso\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.226\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.075\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.088\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.040\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ePLS-A\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.505\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.235\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.237\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.165\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.231\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.191\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\\\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eSpatial Wealth Distribution and Inequality Analysis in Africa\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis analysis utilizes the African settlements dataset released by the Humanitarian OpenStreetMap Team. The original dataset contained 284,968 records, from which entries labeled \u0026quot;place=village or suburb or town\u0026quot; were selected. Points within a 2-kilometer radius were spatially aggregated using the DBSCAN clustering algorithm to ensure corresponding imagery did not overlap. Remote sensing imagery centered on each settlement was acquired using Google Earth Engine (GEE), ultimately yielding 59,998 valid image samples.\u003c/p\u003e\n\u003cp\u003eWealth estimates were generated for each village using the trained model. Within a 2.5 km radius around each settlement, the mean wealth and Gini coefficient were computed to assess local wealth levels and inequality.\u003c/p\u003e\n\u003cp\u003eDue to variations in sampling and spatial coverage of OSM settlements\u0026mdash;such as denser populations, higher infrastructure levels, and more frequent OSM annotations in West Africa\u0026mdash;the 59,998 settlements were divided into four geographic regions: North Africa (north of the Sahara), West Africa, East Africa, and Southern Africa. This classification enables comparative analysis of spatial wealth and inequality across regions. West Africa accounts for the largest proportion of settlements (54%), followed by East Africa (31%), Southern Africa (8%), and North Africa (7%). The regional division is informed by geographic, institutional, and socioeconomic factors, with detailed criteria provided in Supplementary Table S6.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCalculating Population Distribution by Wealth Tiers\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe wealth map was resampled to 100-meter resolution, and we combined the map with WorldPop (100m) [26] and the 2019 World Settlement Footprint (WSF2019) dataset to estimate the population size living within different wealth quantiles. WorldPop only counts residential regions filtered by the WSF2019 mask. The 20th and 80th percentiles were calculated at the national level. Based on these values, boundaries were established to delineate relatively underdeveloped regions (\u0026le;20%) and developed regions (\u0026ge;80%), with the remaining population categorized as ordinary regions (20%\u0026ndash;80%). To ensure consistency with official statistics, we further multiplied the proportion by the total national population as reported by the World Bank in 2020, thereby deriving the final stratified population estimate.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eData availability\u003c/p\u003e\n\u003cp\u003eThe data product has been deployed on GEE and is publicly accessible at https://code.earthengine.google.com/ee8b591dc66c88a7158e3df82db911ea\u003c/p\u003e\n\u003cp\u003eCode availability\u003c/p\u003e\n\u003cp\u003eCode to replicate all findings in the paper are available at https://github.com/UsersLab-tx/Africa_Wealth_Mapping\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAlkire, Sabina, and Maria Emma Santos. 2014. \u0026ldquo;Measuring Acute Poverty in the Developing World: Robustness and Scope of the Multidimensional Poverty Index.\u0026rdquo; World Development 59:251\u0026ndash;274. https://doi.org/10.1016/j.worlddev.2014.01.026.\u003c/li\u003e\n \u003cli\u003eLiu, Qi, Lei Gao, Zhaoxia Guo, Yucheng Dong, Enayat A. Moallemi, Sibel Eker, Jing Yang, Michael Obersteiner, and Brett A. Bryan. 2023. \u0026ldquo;Robust Strategies to End Global Poverty and Reduce Environmental Pressures.\u0026rdquo; One Earth 6 (4): 392\u0026ndash;408. https://doi.org/10.1016/j.oneear.2023.03.007.\u003c/li\u003e\n \u003cli\u003eThomson, D.R.; Kuffer, M.; Boo, G.; Hati, B.; Grippa, T.; Elsey, H.; Linard, C.; Mahabir, R.; Kyobutungi, C.; Maviti, J.; et al. Need for an Integrated Deprived Area \u0026ldquo;Slum\u0026rdquo; Mapping System (IDEAMAPS) in Low- and Middle-Income Countries (LMICs). Soc. Sci. 2020, 9, 80. https://doi.org/10.3390/socsci9050080\u003c/li\u003e\n \u003cli\u003eKuffer, M., Wang, J., Nagenborg, M., \u0026amp; Pfeffer, K. (2018). The scope of earth-observation to improve the consistency of the SDG slum indicator. \u003cem\u003eISPRS International Journal of Geo-Information, 7(11), 428.\u003c/em\u003eDOI:10.3390/ijgi7110428\u003c/li\u003e\n \u003cli\u003eYeh, C., Perez, A., Driscoll, A. et al. Using publicly available satellite imagery and deep learning to understand economic well-being in Africa. Nat Commun 11, 2583 (2020). https://doi.org/10.1038/s41467-020-16185-w\u003c/li\u003e\n \u003cli\u003eTusting, L. S. et al. Mapping changes in housing in sub-Saharan Africa from 2000 to 2015. Nature 568, 391\u0026ndash;394 (2019).\u003c/li\u003e\n \u003cli\u003eOsgood-Zimmerman, A. et al. Mapping child growth failure in Africa between 2000 and 2015. Nature 555, 41 (2018).\u003c/li\u003e\n \u003cli\u003eFrostad, J., Nguyen, Q., Baumann, M. M., Blacker, B., Marczak, L., Deshpande, A., Wiens, K., et al. Mapping development and health effects of cooking with solid fuels in low-income and middle-income countries, 2000\u0026ndash;18: a geospatial modelling study. Lancet Glob. Health 10, e1522\u0026ndash;e1535 (2022). DOI:10.1016/S2214-109X(22)00332-1\u003c/li\u003e\n \u003cli\u003eGraetz, N. et al. Mapping local variation in educational attainment across Africa. Nature 555, 48 (2018).\u003c/li\u003e\n \u003cli\u003eTatem AJ, Gething PW, Bhatt S, Weiss D and Pezzulo C (2013) Pilot high resolution poverty maps, University of Southampton/Oxford DOI: 10.5258/SOTON/WP00127\u003c/li\u003e\n \u003cli\u003eUN‐Habitat. (2004). The challenge of slums: global report on human settlements 2003. \u003cem\u003eManagement of Environmental Quality: An International Journal\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(3), 337-338.\u003c/li\u003e\n \u003cli\u003eSahn, D. E. \u0026amp; Stifel, D. Exploring alternative measures of welfare in the absence of expenditure data. \u003cem\u003eRev. Income Wealth\u003c/em\u003e\u003cstrong\u003e49\u003c/strong\u003e, 463\u0026ndash;489 (2003).\u003c/li\u003e\n \u003cli\u003eFilmer, D. \u0026amp; Pritchett, L. H. Estimating wealth effects without expenditure data-or tears: an application to educational enrollments in states of India. \u003cem\u003eDemography\u003c/em\u003e\u003cstrong\u003e38\u003c/strong\u003e, 115\u0026ndash;132 (2001).\u003c/li\u003e\n \u003cli\u003eDeborah Johnston, Alexandre Abreu, The asset debates: How (not) to use asset indices to measure well-being and the middle class in Africa, \u003cem\u003eAfrican Affairs\u003c/em\u003e, Volume 115, Issue 460, July 2016, Pages 399\u0026ndash;418, https://doi.org/10.1093/afraf/adw019\u003c/li\u003e\n \u003cli\u003eSahn, D. E., \u0026amp; Stifel, D. C. (2000). \u003cem\u003eAssets as a measure of household welfare in developing countries\u003c/em\u003e. Center for Social Development Research Report. Washington University in St. Louis.https://openscholarship.wustl.edu/cgi/viewcontent.cgi?article=1025\u0026amp;context=csd_research\u003c/li\u003e\n \u003cli\u003eKabudula, C. W., Houle, B., Collinson, M. A., \u0026amp; Kahn, K. (2017). Assessing changes in household socioeconomic status in rural South Africa, 2001\u0026ndash;2013: A distributional analysis using household asset indicators. Social Indicators Research, 133(3), 1047\u0026ndash;1073. https://link.springer.com/article/10.1007/s11205-016-1397-z\u003c/li\u003e\n \u003cli\u003eRutstein, S. O., \u0026amp; Staveteig, S. (2014). \u003cem\u003eMaking the Demographic and Health Surveys Wealth Index comparable\u003c/em\u003e.\u003c/li\u003e\n \u003cli\u003eResce, G. (2021). \u003cem\u003eWealth-adjusted Human Development Index\u003c/em\u003e. \u003cem\u003eJournal of Cleaner Production\u003c/em\u003e.\u003c/li\u003e\n \u003cli\u003eAlkire, S., Conconi, A., \u0026amp; Seth, S. (2014). \u003cem\u003eMultidimensional Poverty Index\u003c/em\u003e\u003cem\u003e\u0026nbsp;2014: Brief methodological note and results\u003c/em\u003e. Oxford Poverty and Human Development Initiative (OPHI).\u003c/li\u003e\n \u003cli\u003eSmits, J., Steendijk, R. The International Wealth Index (IWI). Soc Indic Res 122, 65\u0026ndash;85 (2015). https://doi.org/10.1007/s11205-014-0683-x\u003c/li\u003e\n \u003cli\u003eSamper, J.; Shelby, J.A.; Behary, D. The Paradox of Informal Settlements Revealed in an ATLAS of Informality: Findings from Mapping Growth in the Most Common Yet Unmapped Forms of Urbanization. Sustainability 2020, 12, 9510.\u003c/li\u003e\n \u003cli\u003eLee, K., \u0026amp; Braithwaite, J. (2022). High-resolution poverty maps in Sub-Saharan Africa. World Development, 159, 106028. https://doi.org/10.1016/j.worlddev.2022.106028\u003c/li\u003e\n \u003cli\u003eMetzger, N., Daudt, R. C., Tuia, D. \u0026amp; Schindler, K. High-resolution population maps derived from Sentinel-1 and Sentinel-2. Remote Sens. Environ. (2024).\u003c/li\u003e\n \u003cli\u003eMetzger, N., Vargas-Mu\u0026ntilde;oz, J. E., Daudt, R. C., Kellenberger, B., Whelan, T. T. T., Ofli, F., ... \u0026amp; Tuia, D. (2022). Fine-grained population mapping from coarse census counts and open geodata. Scientific Reports, 12(1), 20085.\u003c/li\u003e\n \u003cli\u003eJacobs, N., Kraft, A., Rafique, M.U., Sharma, R.D., 2018. A weakly supervised approach for estimating spatial density functions from high-resolution satellite imagery. In: SIGSPATIAL International Conference on Advances in Geographic Information Systems. pp. 33\u0026ndash;42.\u003c/li\u003e\n \u003cli\u003eTatem, A. J. (2017). WorldPop, open data for spatial demography. Scientific Data, 4, 170004. https://doi.org/10.1038/sdata.2017.4\u003c/li\u003e\n \u003cli\u003eLuo, E., Kuffer, M., \u0026amp; Wang, J. (2022). \u003cem\u003eUrban poverty maps \u0026ndash; From characterising deprivation using geo-spatial data to capturing deprivation from space\u003c/em\u003e. \u003cem\u003eSustainable Cities and Society, 84\u003c/em\u003e, 104033. https://doi.org/10.1016/j.scs.2022.104033\u003c/li\u003e\n \u003cli\u003eKarsai, M\u0026aacute;rton et al. \u0026ldquo;A Comparative Analysis of Wealth Index Predictions in Africa between three Multi-Source Inference Models.\u0026rdquo; \u003cem\u003eArXiv\u003c/em\u003e abs/2408.01631 (2024): n. pag.\u003c/li\u003e\n \u003cli\u003eGorelick, N., Hancher, M., Dixon, M. et al. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment, 202, 18\u0026ndash;27. https://doi.org/10.1016/j.rse.2017.06.031\u003c/li\u003e\n \u003cli\u003eFilippucci, P., Pulvirenti, L., Chini, M. et al. (2020). Sentinel-1 for mapping floods at the scale of the European continent. Remote Sensing, 12(2), 211. https://doi.org/10.3390/rs12020211\u003c/li\u003e\n \u003cli\u003eMicrosoft. (2023). \u003cem\u003eGlobal ML Building Footprints\u003c/em\u003e. GitHub. Retrieved April 15, 2025, from https://github.com/microsoft/GlobalMLBuildingFootprints\u003c/li\u003e\n \u003cli\u003eEssential background in Pesaresi, M. et al. (2024) \u0026quot;Advances on the Global Human Settlement Layer by joint assessment of Earth Observation and population survey data\u0026quot;, International Journal of Digital Earth, 17(1).\u003c/li\u003e\n \u003cli\u003eForget, Y., Shimoni, M., Gilbert, M., \u0026amp; Linard, C. (2021). Mapping 20 years of urban expansion in 45 urban areas of sub-Saharan Africa. Remote Sensing, 13(3), 525. https://www.mdpi.com/2072-4292/13/3/525\u003c/li\u003e\n \u003cli\u003eRonneberger, Olaf et al. \u0026ldquo;U-Net: Convolutional Networks for Biomedical Image Segmentation.\u0026rdquo; \u003cem\u003eArXiv\u003c/em\u003e abs/1505.04597 (2015): n. pag.\u003c/li\u003e\n \u003cli\u003eHe, Kaiming et al. \u0026ldquo;Deep Residual Learning for Image Recognition.\u0026rdquo; \u003cem\u003e2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)\u003c/em\u003e (2015): 770-778.\u003c/li\u003e\n \u003cli\u003eAbascal, Angela, et al. \u0026quot;\u0026ldquo;Domains of Deprivation Framework\u0026rdquo; for Mapping Slums, Informal Settlements, and Other Deprived Areas in LMICs to Improve Urban Planning and Policy: A Scoping Review.\u0026quot; \u003cem\u003eComputers, Environment and Urban Systems\u003c/em\u003e, vol. 93, 2022, p. 101770. https://doi.org/10.1016/j.compenvurbsys.2022.101770\u003c/li\u003e\n \u003cli\u003eAbascal, Angela, et al. \u0026quot;Identifying Degrees of Deprivation from Space Using Deep Learning and Morphological Spatial Analysis of Deprived Urban Areas.\u0026quot; Computers, Environment and Urban Systems, vol. 95, 2022, p. 101820. https://doi.org/10.1016/j.compenvurbsys.2022.101820\u003c/li\u003e\n \u003cli\u003eLi, Chengxiu, et al. \u0026quot;Slum and Urban Deprivation in Compacted and Peri-Urban Neighborhoods in Sub-Saharan Africa.\u0026quot; \u003cem\u003eSustainable Cities and Society\u003c/em\u003e, vol. 99, 2023, p. 104863. https://doi.org/10.1016/j.scs.2023.104863\u003c/li\u003e\n \u003cli\u003eThomson DR, Kuffer M, Boo G, Hati B, Grippa T, Elsey H, Linard C, Mahabir R, Kyobutungi C, Maviti J, et al. Need for an Integrated Deprived Area \u0026ldquo;Slum\u0026rdquo; Mapping System (IDEAMAPS) in Low- and Middle-Income Countries (LMICs). \u003cem\u003eSocial Sciences\u003c/em\u003e. 2020; 9(5):80. https://doi.org/10.3390/socsci9050080\u003c/li\u003e\n \u003cli\u003eUNDP (United Nations Development Programme). (2020). Angola Human Development Report 2020. Home | United Nations Development Programme\u003c/li\u003e\n \u003cli\u003eUN Economic Commission for Africa (ECA), \u003cem\u003eThe Economic Implications of the Oil Sector in Sub-Saharan Africa: Angola\u003c/em\u003e, 2015.\u003c/li\u003e\n \u003cli\u003eBeegle, K., \u0026amp; Christiaensen, L. (2019). Accelerating Poverty Reduction in Africa. Washington, DC: World Bank. https://doi.org/10.1596/978-1-4648-1232-3\u003c/li\u003e\n \u003cli\u003eLi, C., Yu, L., Ndugwa, R. \u003cem\u003eet al.\u003c/em\u003e Mapping urban slums and their inequality in sub-Saharan Africa. \u003cem\u003eNat Cities\u003c/em\u003e (2025). https://doi.org/10.1038/s44284-025-00276-0\u003c/li\u003e\n \u003cli\u003eChristiaensen, L., \u0026amp; Todo, Y. (2014). \u003cem\u003ePoverty reduction during the rural\u0026ndash;urban transformation \u0026ndash; The role of the missing middle\u003c/em\u003e. World Development, 63, 43\u0026ndash;58.\u003c/li\u003e\n \u003cli\u003eFosu, A. K. (2017). \u003cem\u003eGrowth, inequality, and poverty reduction in developing countries: Recent global evidence\u003c/em\u003e. Research in Economics, 71(2), 306\u0026ndash;336.\u003c/li\u003e\n \u003cli\u003eBanerjee, A. \u0026amp; Duflo, E. \u003cem\u003ePoor Economics\u003c/em\u003e (2011). Classic evidence on poverty traps and the effectiveness of education/basic-services/SME interventions; your policy framing aligns with these micro-foundations.\u003c/li\u003e\n \u003cli\u003eAlkire, S., \u0026amp; Santos, M. E. (2014). \u003cem\u003eMeasuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index\u003c/em\u003e. World Development, 59, 251\u0026ndash;274. https://doi.org/10.1016/j.worlddev.2014.01.026\u003c/li\u003e\n \u003cli\u003eSatterthwaite, D. (2017). \u003cem\u003eThe impact of urban development on risk in sub-Saharan Africa\u0026rsquo;s cities with a focus on small and intermediate urban centres\u003c/em\u003e. International Journal of Disaster Risk Reduction, 26, 16\u0026ndash;23. https://doi.org/10.1016/j.ijdrr.2017.09.025\u003c/li\u003e\n \u003cli\u003eWang, L., Long, T., Jiang, W., Adam, E., Wen, C., Jiao, W., \u0026amp; He, G. (2025). Economic well-being assessment: a review of traditional and remote sensing approaches. \u003cem\u003eInternational Journal of Digital Earth\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(1). https://doi.org/10.1080/17538947.2025.2504137\u003c/li\u003e\n \u003cli\u003eEspey, J. et al. Data for development: a needs assessment for SDG monitoring and statistical capacity development. \u003cem\u003eSustain. Dev. Solut. Netw.\u003c/em\u003ehttp://unsdsn.org/wp-content/uploads/2015/04/Data-for-Development-Full-Report.pdf (2015).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8234710/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8234710/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eEconomic well-being data are critical for understanding development conditions and monitoring poverty, yet micro-level socioeconomic conditions in Africa remain poorly understood. Here, we present a high-resolution wealth mapping framework that integrates household survey indicators with spatial wealth proxies, producing the first 10-m economic well-being maps across 30 African countries. Our model explains 74% of the variance (R\u0026sup2;=0.74) and our results reveal that ~\u0026thinsp;36% of the population\u0026mdash;about 340\u0026nbsp;million people\u0026mdash;still reside in the least-developed areas. Expanded analysis of nearly 60,000 settlements further shows that Africa remains in a state of \u0026ldquo;equilibrated poverty\u0026rdquo;: while economic growth tends to widen internal disparities, development trajectories follow a nonlinear U-shaped pattern rather than a monotonic trend. Under extreme poverty, inequality actually intensifies, creating compounded risks of poverty and social exclusion. These findings uncover spatial patterns invisible to coarse-scale studies and provide new evidence for poverty alleviation and equitable infrastructure planning across the Global South.\u003c/p\u003e","manuscriptTitle":"High-resolution wealth maps reveal Africa’s nonlinear trajectory of development and inequality","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-11 07:22:55","doi":"10.21203/rs.3.rs-8234710/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c3f8bae1-0bd6-4f70-898f-027aeae87071","owner":[],"postedDate":"December 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":59426979,"name":"Earth and environmental sciences/Environmental social sciences/Sustainability"},{"id":59426980,"name":"Scientific community and society/Geography"}],"tags":[],"updatedAt":"2025-12-11T07:22:56+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-11 07:22:55","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8234710","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8234710","identity":"rs-8234710","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}