A new mapping technique that harnesses big data is being developed to transform the ability of governments and aid organisations to tackle extreme poverty.
Outlined in US journal, Proceedings of the National Academies of Sciences, the poverty maps are created with computational tools developed using a combination of information gathered from cell phones, satellite data and machine learning.
“Despite much progress in recent decades, there are still more than 1 billion people worldwide lacking food, shelter and other basic human necessities,” explained Neeti Pokhriyal, one of the study’s co-lead authors, and a computer science PhD candidate at the University at Buffalo.
In the study, Neeti and Damien Jacques, a PhD candidate at the Universite Catholique de Louvain, Belgium, and the other co-lead, built a poverty map of Senegal.
Two data sets, one put together from billion of calls and texts made by nine million Senegalese mobile phone users, the other from satellite imagery, indicating levels of local infrastructure development and availability of electricity, were analysed by a machine learning platform.
The researchers were then produced maps detailing poverty levels of 552 communities across the country, splitting Senegal in to four regions, and allowing stakeholders to make timely interventions.
And it is a game-changer, as the maps can be updated regularly and at low-cost. Previously, poverty maps would be costly and rely on census and survey data – as and when it is available.
Pokhriyal said the maps can fill in the gaps created between cycles of census taking – and could be applied during periods of conflict and war.