Can a map save lives? Facebook thinks so.

Using satellite photos and artificial intelligence (AI), the company has created a high-resolution map of population density in Africa that it says is more detailed than any of its predecessors.

To create the map, publicly available images were broken up into billions of 30-by-30-metre sections, and each was analysed to assess how many people might live there.

Facebook says it eventually wants to map the whole world’s population.
Image: Facebook

There are more than a billion people across Africa and population densities vary by region and country. The overall figure for the continent is 45 people per square kilometre, and more than 40% of the population is based in urban centres.

That means hundreds of millions of people live in rural, sometimes remote, locations. And in countries like Rwanda, where only a quarter of roads are paved or surfaced, remote can also mean inaccessible or even hidden.

The project mapped hundreds of millions of buildings to find where people live.
Image: Facebook

The billions of blocks into which the satellite images are divided each have a unique reference. An AI tool assesses each image for the tell-tale signs of human habitation – buildings. The results are then cross-referenced with census data to determine how many people are likely to be living in those areas.

Accurate population density forms arguably the backbone for any public sector or social service intervention you can think of,” says Laura McGorman, a public policy manager with Facebook’s Data for Good division.

For anyone responding to natural disasters, such as the recent cyclones in Mozambique, or epidemics like the Ebola outbreak, delay or confusion around where to direct help or where to find people can have dire consequences.

The Red Cross and the government of Malawi used earlier iterations of Facebook’s maps of Africa as part of a vaccination campaign. Using map data that showed 97% of a particular area was uninhabited, the Red Cross was able to send 3,000 trained volunteers to the areas where they were most needed.

As the maps become more sophisticated they can also be used to help in service development planning. Solar electricity, roads and communication infrastructure, all vital to the ongoing economic development of the region, can be targeted more effectively.