DATASETS
Relative Wealth Index
The Relative Wealth Index predicts the relative standard of living within countries using privacy protecting connectivity data, satellite imagery, and other novel data sources.
Download from HDX
Key Features
Low and middle income country representation
The Relative Wealth Index has been generated for nearly 100 countries worldwide.Granularity
There are roughly 20 million 2.4 km² micro-regions on earth’s surface. This index is an estimate of relative wealth of the people living in each micro-region relative to others in the same country.Nontraditional data sources
This index is built using nontraditional data sources, including satellite imagery and privacy protecting Facebook connectivity data. The index is validated using ground truth measurements from the Demographic and Health Surveys.
Methodology
Step 1: Leverage ground-truth measurements
We leverage measurements of household wealth collected through face-to-face surveys conducted by the United States Agency for International Development with 1,457,315 unique households living in 66,819 villages in 56 different LMICs around the world.Step 2: Link each village to non-traditional data
We use spatial markers to link villages to non-traditional data sources , including satellite imagery, cellular network data, topographic maps, and privacy protecting connectivity data from Facebook.Step 3: Process data using deep learning
We process the non-traditional data using deep learning and other computational algorithms, which convert the raw data to a set of quantitative features for each village.Step 4: Train supervised machine learning model
We use these features to train a supervised machine learning model that predicts the relative wealth of each populated 2.4km2 grid cell on the planet, even in regions where no ground truth data exists.RESOURCES
Using the Relative Wealth Index
This index can be downloaded from Humanitarian Data Exchange.
Download Relative Wealth Index