After Flint’s Crisis, An Algorithm Helps Citizens Find Lead Pipes
It’s been nearly seven years since the beginning of Flint, Michigan’s water crisis, when high levels of lead from corroded lead pipes led to water shortages and health issues for city residents. Since then, many other cities around the country have had their own problems with lead. Researchers estimate that millions of Americans are living with pipes that need to be replaced.
As Wired reported earlier this month, Toledo, Ohio is one of the latest cities trying to get ahead of its legacy of lead plumbing, with the help of an algorithm created by University of Michigan researchers. The model was originally created to help the city of Flint more quickly—and less expensively—target which homes were most likely to need their pipes replaced.
The same researchers are now working as a private company, called BlueConduit, to help other cities do the same work. And in Toledo, they’re working in close partnership with the city and community organizations.
Ira talks with University of Michigan professor and BlueConduit co-founder Eric Schwartz, and Alexis Smith of the nonprofit Freshwater Future, about the work ahead for Toledo, and why deploying an algorithm effectively depends on community trust and input.
Curious if your own water pipes contain lead? EPA-funded project Crowd The Tap has a free tutorial for finding your water service line—and determining the materials of your pipes. The organization’s mission is to ensure safe drinking water in the United States. By sharing what you observe, you can help identify areas for tap water testing and infrastructure replacement. Learn about your pipes, and how you can help at CrowdTheTap.org
Invest in quality science journalism by making a donation to Science Friday.
Alexis Smith is a community and technology associate at Freshwater Future in Toledo, Ohio.
Eric Schwartz is a professor in the Ross School of Business at the University of Michigan and co-founder of BlueConduit in Ann Arbor, Michigan.
IRA FLATOW: We’ve been hearing about the water crisis in Flint. But Flint isn’t the only city in the nation with lead pipes. Millions of Americans are estimated to be exposed to lead-tainted water, and there is no safe level of exposure, according to the CDC and the EPA. But cities have been slow to replace pipes. It is an expensive process that also relies on having accurate data about which homes need it. I mean, you don’t want to dig $1,000 hole in a home that already has safe copper pipes, right?
Well, enter machine learning, an idea that a team of researchers at the University of Michigan first had when looking at the task ahead in Flint. It offered a solution– an algorithm that could take what cities do know about each home and calculate the probability of lead pipes for each.
The result– a ranked list of where to start digging to have the most immediate benefit. And now they’re offering the same algorithm to other cities around the country like Toledo, Ohio. Speaking of which, Alexis Smith is an associate at Freshwater Future, a nonprofit that works with communities around the Great Lakes Region on access to clean water. She joins us from Toledo. Welcome, Alexis.
ALEXIS SMITH: Thank you. Hi. Pleasure to be here.
IRA FLATOW: Nice to have you. Dr. Eric Schwartz is a professor in the Ross School of Business at the University of Michigan in Ann Arbor and a member of the team behind the algorithm and a co-founder of the company Blue Conduit. Welcome, Eric.
ERIC SCHWARTZ: Hi, Ira. Thanks for having me.
IRA FLATOW: You’re quite welcome. Alexis, the country is full of aging infrastructure, including those aging pipes. What’s the scope of the problem in Toledo specifically?
ALEXIS SMITH: Sure, yeah. So I’ll tell you that there are 130,000 lead service line connections still in Toledo. And that’s according to the city of Toledo’s estimate, which can be found on their website. 30,000 of those are city-owned. And so from that, there’s probably 25% of what’s left that are privately owned and therefore impacting families.
IRA FLATOW: And now in 2021, your organization, Freshwater Future, is partnering with Eric’s company. And Toledo is working on getting those pipes out. How much progress has the city made so far?
ALEXIS SMITH: Well, I would say that they’ve focused majority of their resources right now on the city side, while offering loan installments and things like that to residents who want to change the private side. However, the issue with just offering loans and not being a little bit more proactive and aggressive with our approach is that the areas where these lines are most concentrated are in areas where families are already despaired by other things– are low-income neighborhoods. They’re going to have issues repaying that loan. And like these repairs can be anywhere from $3,000 to $10,000 replacements. It’s almost as if they’re backed up into a corner, where they have these lead service lines, but there’s not much support in replacing them.
IRA FLATOW: Eric, so you come up with this algorithm which really, you put the data in. It makes decisions about where to look, where not to look. How successful is the algorithm when you apply it?
ERIC SCHWARTZ: Well, in practice what we’re doing is really– we’re providing a characterization of the uncertainty to the ultimate decision-makers, who are the communities and water utilities themselves. By providing those lists, we’re saying, as you had mentioned, a rank ordering from the homes that are most likely to have lead, all the way down to the least likely, perhaps even by block or by neighborhood. However, the groups on the ground actually going and doing this work are going to be organizing so that when they go, they’re going to get as much lead as possible, given that they are going to spend that money and send crews and close down the streets and get permits.
And so success in that regard we really measure by what we call a “hit rate.” So for every 100 times that those shovels or equipment is going in the ground, how many of those actually are finding lead? And so in Flint, we were running around 80% of all attempts, we’re finding lead. In other cities, now we’ve seen that can be even higher.
IRA FLATOW: Alexis, you’re living in one of the communities most affected by the lead in this city. Does that percentage sort of meet– match what you know about the homes?
ALEXIS SMITH: I would say it does, especially since this is a historical community that has not had any really updated infrastructure in quite some time or any sort of investment. For example, the house that I live in right now has been in my family for generations. You know, this house itself, outside of the work that we’ve done ourselves, has probably had little infrastructure updates. And so if I can say that, I can pretty much guarantee that for my neighbors, who’ve been here for a very long time. And so the short answer is, yes, it is likely that we have a very concentrated lead service line population in this area.
IRA FLATOW: And how do your neighbors feel? Are they happy that they have to wait for the inspections and replacements, knowing that at least they’ll be coming? Or are they frustrated by the wait?
ALEXIS SMITH: You know, on top of everything that we’re seeing and experiencing with this pandemic, I think the initial response is frustration. But as we have these conversations, and we talk about how this is not something that we’re just planning to have done to us, we’re doing this together. And we want the community to be a part of every step of that process to make sure that the historical information that is needed to make this as accurate as possible is there.
This is going to take them being involved to really be ultimately successful. And so going through and having those conversations definitely provides relief. Of course, the initial response is frustration because these families are not just dealing with water issues.
We’re dealing with living in a food desert. We’re dealing with the vacancies. We’re dealing with having poorly lit streets and not having enough attention and investment being poured into our neighborhoods. And so there is understandably more frustration, but relief in knowing that not only is there a solution, but there is a citizen-driven, partnership-supported solution.
IRA FLATOW: Eric, how important from the science side is community participation? Why can’t you just roll into town, hand over a list of likely lead pipes, and be done with it?
ERIC SCHWARTZ: One way that combining the data science and machine learning and statistics with community perspective is we want to make sure that we are addressing priorities. So what Alexis was just talking about, and something that we’ve talked to Freshwater Future and Toledo about, is not just providing a list where they should go to dig and replace the pipes, but also, based on the predicted likelihood of lead service lines, where should they really concentrate certain kinds of public communication efforts, and really just the efficiency– to make sure. These are going to be really limited funds.
Everyone wants to get their lead pipes replaced. They also want to know if they even have lead Pipes but the next month, the next three months, the next year, not everyone is going to be able to get all of that taken care of in every city. And so the question is, how do you do this and allocate resources equitably, not in a very top-down, algorithmic way?
IRA FLATOW: This is Science Friday from WNYC Studios. There was a news item that may have snuck past a lot of people late last year, and that was that the EPA has updated its lead and copper rule for the first time since 1991 and added some new requirements for cities to understand and act on their lead pipes. Will this change, Eric? Make a meaningful difference in people’s lives?
ERIC SCHWARTZ: The new lead and copper rule has not gone as far as a lot of advocates, and us included, would have liked. One change that I would like to highlight that is a positive is there is now a enforceable requirement that has every water system above a certain size making public the materials of service lines for every address, even if it means saying for these 10,000 addresses, we don’t know. In fact, there’s a requirement that you have to announce to those residents that you don’t know the material of their service line, or that you do know it, and it is lead and needs to be replaced.
And so what that’s doing is creating just the right amount of pressure for communities to ask the question and for utilities to have to answer and say, well, here’s our best answer. And here’s the investment we’re going to make to go out and prioritize which homes, not just to replace, but also where to look. And that’s something that we are really excited about– taking these best practices and statistics to communicate these best practices in this community engagement from Flint or Toledo to other places because it allows us to really better characterize all this uncertainty for the utility to better make their decisions and better communicate with their constituents and customers.
IRA FLATOW: Alexis, one last question for you, because I know when you have old cities, old pipes, old water problems, the water problems just don’t end with lead, do they? There are a lot of other things going on.
ALEXIS SMITH: Well, one thing I will say is that back in 2014, Toledo had its own water crisis when we had toxic algal blooms that made our water undrinkable, untouchable, really. We were without water for three days here in Toledo. The distrust is there, and it exists. There are still people going out, buying bottled water. There’s still people who don’t trust drinking the water, even against all the city’s reassurance.
And so when we talk about those toxins, when we talk about the contamination, we’re also getting into another issue. And really, that’s going to be the overarching challenge for not just the Great Lakes states to face, but the entire country– water affordability. As these contaminants come into our systems, and it becomes more expensive and more complex to filter out these contaminants, the bill for reckoning with that increases.
We had a multi-million dollar system installed to filter out these algal blooms. And how are we paying for that? Well, we’re raising water rates for residents. Shutoffs are going to be an issue. When we lead to shutoffs, we lead to a whole other public health issue.
Especially in the middle of a pandemic, we need to have access to water. We saw that resoundingly in Detroit. There are tons and tons of water issues that need to be addressed, and lead is just one of them.
IRA FLATOW: And that’s something we’ll have to wait to talk about later, because as we all know, water is the issue of the future. Thank you so much to my guests, Dr. Eric Schwartz, professor at the University of Michigan in Ann Arbor, co-founder of Blue Conduit; and Alexis Smith, community and technology associate for the non-profit organization Freshwater Future, based in Toledo. Thank you both.
ALEXIS SMITH: Thank you for having me.
ERIC SCHWARTZ: Thank you.
IRA FLATOW: And if our listeners are curious about what their water pipes are made of, you can find out. Learn how to tell if your pipes are lead, copper, steel, or PVC, and help identify areas for tap water testing and infrastructure replacements. Want to know how to do all of that? Well, head to our crowdsourced science project, “Crowd the Tap.” We have a link for it up on our website sciencefriday.com/leadpipes.