Machine learning tower detection and mapping work
At the beginning of 2018 Development Seed and the World Bank ran a project of identification on high-voltage, infrastructure for developing countries, Pakistan, Zambia, and Nigeria.
The strategy was, running an ML to identify tower across of those 3 countries and then the Data team jump to map those places, this was a really challenging work, making the organization of the resources in ML and on the Mapping side.
The Data team completed the work in around 2,540,099 km² in approximately 132 days, editing 131,832 objects in OpenStreetMap.
You can always read more about the documentation of Mapping the electric grid.