Modeling high resolution spatial poverty distribution using satellite images

Abstract

In developing countries, despite the best efforts of Government bodies and the international development community, accurate estimates of poverty and economic welfare in vast area remain rare. The major reason for this large data deficit is due to large gaps in the availability of household data measuring economic welfare parameters. Even the accuracy of the predictions made from data available is always questionable due to small sample size for the population for which surveys are being carried out. The primary reasons for these problems can be attributed to huge time span between consecutive surveys, high cost associated with carrying out surveys at large scale or nonaccessibility of certain areas to carry out the surveys.