Listen to our co-Founder and co-CEO, Eric Schwartz, talk to Paul Gagliardo MPH, P.E. on The Water Entrepreneur Podcast. Paul and Eric covered BlueConduit’s origins, the challenges of using only home age to predict lead service lines, and the skill sets that make academics good entrepreneurs and business managers good academics.
“At the end of the day we want to be digging and prioritizing to dig where the lead is and dig where it’s going to make the biggest impact on public health.”
Eric Schwartz (26:55 – 27:10)
What We’ve Learned
You can’t use year of home alone. It is helpful, the year of home of course is helpful. There’s a lot of reasons why we have to be careful about it.
For instance, building patterns are strange. Trends went in many different directions….It’s not linear, either.The other issue is the data itself are not really the data about when the service line was installed. It’s “year built,” typically, and year built can mean the year that the new home was built after the old home was torn down. So you could have, sometimes, year 1993, and a lead pipe in there, and sure the lead pipe was installed in ’62, and the home was knocked down and rebuilt in the ’90s.
The data is tricky there and so you don’t want to depend only on that, just like you don’t want to depend largely on records that were largely written eighty years ago.
Eric Schwartz (24:55-25.57)
About BlueConduit
BlueConduit originated the approach of using machine learning to predict lead pipes and have been doing it longer than anyone else.
Read about our origins in Flint: BlueConduit’s data science predictions of drinking water service line materials improved planning, reduced cost, and provided replacement priorities based on risk of lead exposure.
Listen to a recent interview of Eric’s with Waterloop where he explains the relationship between A.I., machine learning, predictive modeling and lead pipes.
