Mapping poverty

9 Wrapping up

This guide has provided a practical example of how to apply Small Area Estimation (SAE) methods using geospatial data, with a specific focus on estimating poverty at the admin 3 level in Northern Malawi. We have covered each step of the process—from setting up the R environment and handling vector and raster data, to preparing survey datasets, engineering spatial features, estimating SAE models, and mapping the final results. While the example used here is for Malawi, the overall approach is designed to be generalisable. With suitable data, the same methods can be applied in a wide range of country contexts and for different SDG-related indicators.

We note that we are not suggesting that the methods and data used here should always be a practitioner’s starting point for SAE. For example, if recent census data is available, we would encourage practitioners to instead consider using methods appropriate for census data – e.g. household-level models – instead of the sub-area models we covered here. It is important to explore all possible available data and choose methods appropriate for those data.

We also encourage users to read through other available resources, including the SAE4SDGs wiki and the Primer on Small Area Estimation with Geospatial Data.