Subnational Poverty Models

Granular Data for Actionable Intelligence

National poverty forecasts do not tell the entire story. While an entire country could be on-track to achieve Sustainable Development Goal 1 by 2030, certain regions might be left behind. For more targeted and granular data, subnational models are necessary.

 

Pakistan Subnational Poverty Model

In partnership with the International Growth Centre (IGC), we developed a subnational poverty model for Pakistan incorporating multiple data sources and nighttime imagery. Using the latest data science techniques, World Data Lab was able to produce one of the first consistent and accurate subnational poverty models of the Indian subcontinent providing decision-makers and researchers with regional, granular, and actionable data.

Brazil Subnational Poverty Model

Along with our local Brazilian affiliate, we developed a subnational poverty model for Brazil. The results from this model will be featured on the World Poverty Clock. The benefit of subnational data is that it provides more regional and granular information than a national forecast.

North Korea Poverty Model

No country’s economic condition is more difficult to model than North Korea’s. With almost no publicly available data sources, traditional econometric methods can not be used. Using a variety of nighttime satellite imagery and land-use data along with new “twinning” techniques for comparing similar regions, World Data Lab has pioneered the creation of the world’s first-ever subnational income and poverty model of North Korea in partnership with the NOMIS Foundation, the International Institute for Applied Systems Analysis (IIASA) and the Vienna University of Economics and Business. The results of our projections have been published in The Economist.