Centaur VGI: Evaluating a human-machine workflow for increased productivity during humanitarian mapping campaigns

The spatial and temporal distribution of global map data is highly unequal, with large areas of the world suffering from a paucity of data. Volunteered geographic information (VGI) has been vaunted as a potential solution, but is also criticised for reinforcing rather than alleviating inequalities. Human- machine workflows have been suggested to improve the speed and quality of VGI production for poorly mapped regions, but this ability is yet to be fully evaluated. This paper provides the first detailed evaluation of a human-machine workflow, testing its ability to produce high quality, timely data in remote regions often neglected by humanitarian mapping campaigns.