GRID3 — Sierra Leone's AI-Powered Population Mapping for Vaccination Micro-Planning

Sierra Leone's last full national census was in 2015, leaving planners with an outdated picture of where people live. Under the GRID3 partnership, Statistics Sierra Leone works with Esri, Maxar, Fraym and the Global Partnership for Sustainable Development Data to combine satellite imagery, machine-learning building-footprint extraction and household survey data into population, age and sex estimates for every 1-hectare cell in the country. Starting in 2021 with COVID-19 outreach, the Ministry of Health and Sanitation's Expanded Programme on Immunization paired this data with school-enrolment records to build microplanning maps, producing 223 such maps for the October 2022 HPV vaccination campaign run with the Clinton Health Access Initiative, Gavi and WHO.

Against a projected target of 153,991 ten-year-old girls, the campaign administered 177,141 HPV vaccine doses in a single week (3-11 October 2022) — about 115% of the modelled target population. Sierra Leone has since run a larger, multi-age-cohort national HPV campaign that pushed national coverage above 70%, though that later campaign's reporting does not specify whether GRID3 maps were reused.

The underlying population counts are machine-learning estimates calibrated against a decade-old census and periodic surveys rather than a fresh headcount, so accuracy still depends on regular ground-truthing; a single campaign's success also reflects overall programme design — cold-chain logistics, community mobilisation, school partnerships — as much as the mapping technology itself.