Satellite Image Analysis for Poverty Prediction

High-resolution satellite imagery is increasingly available globally and contains abundant information about landscape features that could be correlated with economic activity. Unfortunately such data are highly unstructured and thus challenging to extract meaningful insights from at scale even with intensive manual analysis. Recent applications of deep learning techniques to large-scale image data sets have led to marked improvements in fundamental computer vision tasks such as object detection and classification but these techniques are generally the most effective in supervised learning.