Combining satellite imagery and machine learning to predict poverty

By Neal Jean, Marshall Burke, Sang Michael Xie, W. Davis, David B. Lobell, Stefano Ermon. 2016

1,609 citations recorded by the source corpus.

Subfields: Global and Planetary Change, Pollution

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Global and Planetary Change

1. H. Haberl
2. Yang Zhou
3. Yan-sui Liu
4. K. Erb
5. Anuj Karpatne
6. Yurui Li
7. David B. Lobell
8. Vipin Kumar

In Pollution

1. F. Krausmann
2. Christoph Görg
3. Sendhil Mullainathan
4. Dominik Wiedenhofer
5. H. Haberl

Citations in the corpus

• WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, positive): “…efforts in applications such as tracking deforestation (Hansen et al., 2013), population density mapping (Tiecke et al., 2017), poverty mapping (Abelson et al., 2014; Jean et al., 2016), crop yield prediction (Wang et al., 2020), and other economic tracking applications (Katona et al., 2018).”
• Targeted poverty alleviation and land policy innovation: Some practice and policy implications from China (2018, background): “…food production and poverty reduction (Fischer and Hajdu, 2015), the anti-poverty effect of economic input (Ravallion, 2001; Ravallion and Chen, 2007; Montalvo and Ravallion, 2010; Barrett and Carter, 2013; Lü, 2015; Gao et al., 2015) and mapping of poverty (Jean et al., 2016).”
• Machine Learning for the Geosciences: Challenges and Opportunities (2017, background): “been explored in [97], [98] for mapping poverty using high-”
• Calibrated ensembles can mitigate accuracy tradeoffs under distribution shift (2022, background): “Similarly, when predicting poverty from satellite imagery, models trained on data from a few countries work poorly on new countries, particularly those where labels are scarce due to resource constraints [Jean et al., 2016].”
• Semantic segmentation of slums in satellite images using transfer learning on fully convolutional neural networks (2019, neutral): “However, some forms of urban poverty in particular can be directly associated with the morphology of the built environment, though (Sandborn and Engstrom, 2016; Jean et al., 2016; Wurm and Taubenböck, 2018).”
• Machine Learning and Deep Learning frameworks and libraries for large-scale data mining: a survey (2019, background): “involving images such as poverty prediction from satellite imagery (Jean et al. 2016).”
• The Material Stock–Flow–Service Nexus: A New Approach for Tackling the Decoupling Conundrum (2017, background): “Mapping of human activities is so far largely restricted to population density maps and other datasets mainly derived from proxies such as nighttime lights [97,127,150].”
• Poverty from Space: Using High Resolution Satellite Imagery for Estimating Economic Well-being (2017, neutral): “Previous researchers [7] have employed a transfer learning”
• Deep Learning Microscopy: Enhancing Resolution, Field-of-View and Depth-of-Field of Optical Microscopy Images Using Neural Networks (2017, methodology): “CNNs form a rapidly growing research field with various applications in e.g., image classification 3 , annotation 4 , style transfer 5 , compression 6 , and deconvolution in photography 7–9 , among others 10–13 .”
• Functional Map of the World (2017, background): “Recently, there have been several, mostly successful, attempts to leverage techniques that were founded on first-person imagery and apply them to remote sensing data [15, 21, 30].”
• Gamut: A Design Probe to Understand How Data Scientists Understand Machine Learning Models (2019, background): “With recent advances in machine learning (ML) [29, 37, 58, 65], people are beginning to use ML to address important societal problems like identifying and predicting cancerous cells [14, 32], predicting poverty from satellite imagery to inform policy decisions [27], and locating buildings that are s”
• Using deep learning and Google Street View to estimate the demographic makeup of neighborhoods across the United States (2017, neutral): “The model could also be improved by integrating other types of data, such as satellite images (22), social networks (4), and economic data pertaining to local consumer behavior in particular geographic regions.”