Earth Observations and Machine Learning for Gridded Macroeconomic Data

Paper by Marco Marini, Jim Tebrake, and Andinet WoldemichaelThis paper presents a methodology for downscaling official national and subnational macroeconomic data into high-resolution grids using spatial machine learning techniques. Traditional macroeconomic data are highly aggregated and obscure the spatial distributions needed to understand and quantify local economic activity, impacts of physical risks, and infrastructure gaps. Our hierarchical approach integrates official macroeconomic accounts with Earth observation predictors, such as nighttime lights, built-up areas, land cover, and gridded population, to deliver high-resolution, gridded estimates that remain fully consistent with official subnational and national totals. The empirical application focuses on constructing experimental gridded GDP by ten-sector industry for Canada and the United States. The methodology complements and enhances official statistics by adding spatial granularity through open geospatial datasets and can be extended to generate gridded capital stock estimates…(More)”.