Two years ago, Facebook launched a project that aims to help developing countries obtain real data on population densities by building in the African continent.
More specifically, the Boston-based Facebook AI team divided satellite images into grids of 30x30 squares and then used Machine Learning techniques to analyze each square for evidence of any human settlement there. Once the images were analyzed and it was possible to see where homes were located, these datasets were combined with those of the available census to find out how many people could be living in the analyzed homes.
If the core idea was to find out where exactly the population was concentrated (the concentration of population by territories was known, but not by buildings), with this project they had finally succeeded.
Some public health organizations have already spoken out and highlighted how this valuable data will be used for fumigation campaigns against malaria, for example.
Other organizations such as Red Cross have managed, together with the Missing Maps Project, to carry out a more strategic campaign against rubella and measles, focusing their efforts at the building level in populated areas and being able to cover more territory with less staff.
Other types of interests aside, the project deserves full recognition as it will help governments in developing countries to tackle certain long-standing problems such as proper resource management to fight diseases, a better distribution of humanitarian aid and an increase in the quality of life of citizens.