In 2012 an Oxford research team began a study in Kenya by fitting 60 village hand pumps with data transmitters. They hoped to monitor the motion of the pump to determine the amount of water extracted on an hourly basis.
If no motion was detected, it meant the pump wasn’t working, and a message was sent to a repair company and workers were sent to fix the problem.
This innovation cut the average repair time from over a month to less than three days.
But in the process of monitoring the pumps, they found a new way to interpret the data from the instruments on the pump handles.
They discovered that when the water is being drawn from a deep aquifer, it produces different vibrations than when the liquid comes from a shallow one. They found by using data from the pumps and rather sophisticated statistical analysis, they could predict levels of groundwater available – valuable information for farming, mining and to communities.