An Algorithm Is Helping a Community Detect Lead Pipes


BlueConduit, an analytics startup, applies statistical models to identify neighborhoods and households that might have lead pipes. The models include dozens of factors, such as the age of the home and proximity of other homes where lead has been found, to help predict likely locations of lead pipes and create a ranking by likelihood that cities can use to prioritize which pipes to examine. In Flint, MI, about 70% of the homes identified using the models had lead pipes, compared to about 15% of homes where excavations did not use the model. The company is working with organizations in dozens of other cities.

Related Stories