
A Faster Fix for Lead Water Lines
Residents in Toledo now have safer water thanks to an ongoing underground renovation. Due to an algae bloom in western Lake Erie in 2014, the city of Ohio received a drinking water advisory.
That raised the issue of water quality and started a campaign to supply the greatest drinking water available. The elimination of lead service lines was at the top of the city’s priority list to do that.
According to Mark Riley, director of the Department of Public Utilities for the city of Toledo, “We realised we had to address lead exposure.” We looked at the partnerships used by Flint, Michigan, in light of the problems there, and they referred us to BlueConduit.
A water analytics startup called BlueConduit created machine learning algorithms to find lead service lines.
Toledo’s lead removal strategy was made possible by Flint’s contamination problems, which also served as inspiration for BlueConduit to create a particular piece of software in the first place.
The Flint catastrophe, according to Eric Schwartz, co-founder of BlueConduit, prompted the question of how many people are contaminated with lead, to which he claims no one had a satisfactory response.
“As a data scientist, I realized this was an opportunity to really help cities like Flint figure out how many lead pipes existed and which homes have lead pipes, and guide the resource allocation so the process can be efficient and equitable,” Schwartz says.
And that’s exactly what BlueConduit did. They helped the city of Flint locate lead pipes and create public-facing maps to help inform the community about replacement.
BlueConduit’s software takes existing data about lead pipes, formulates an inspection plan to verify what is known and unknown and from that, makes predictions of where lead lines may be.
It provides real-time maps so water utility managers and consumers can see progress on lead pipe replacements in their community.
Getting started
The software was an obvious choice for Riley and the city of Toledo, and they have now been working with BlueConduit for over two years.
The beginning stages were all about building a base of information. “They began by collecting data on all of our service lines, especially documentation on lead lines,” Riley says.
“They wanted to see what type of pipe material we had from the city side as well as the pipes going into the customer’s home.”
Cities are urged to upload any data they have about their service lines and parcels so from the very beginning, the software can build from what is known, like how many homes, and which homes have actually had eyes on material in recent years.
Gathering as much preexisting data as possible is crucial to the process. The more data city officials could provide from the start, the more accurate predictive models would become.
“BlueConduit helped our staff understand what sort of information was needed and how to transfer it,” Riley says. “They were very straightforward and thorough, asking a lot of questions to make sure they understood how we keep our city data.”
The beginning stages were all about building a base of information. “They began by collecting data on all of our service lines, especially documentation on lead lines,” Riley says. “They wanted to see what type of pipe material we had from the city side as well as the pipes going into the customer’s home.”
Cities are urged to upload any data they have about their service lines and parcels so from the very beginning, the software can build from what is known, like how many homes, and which homes have actually had eyes on material in recent years. Gathering as much preexisting data as possible is crucial to the process.
The more data city officials could provide from the start, the more accurate predictive models would become. “BlueConduit helped our staff understand what sort of information was needed and how to transfer it,” Riley says.
“They were very straightforward and thorough, asking a lot of questions to make sure they understood how we keep our city data.”
After collecting available data, BlueConduit created early models that predicted areas with the highest probability of lead service lines.
From those early predictive models, city officials were able to do curbside inspections of homes to see firsthand and confirm what type of material was being used.
“We could then plan our lead service line replacement program to first target the neighborhoods with the greatest concentration of lead,” Riley says.
The models are constantly being updated with new data as officials replace pipes and visit more areas of the city for inspections, making them more and more accurate over time.
Data collected, as well as the corresponding maps and models, are not only available to city officials but to anyone interested. Riley welcomes and encourages the public to stay engaged, so maps are available in a public online portal.
“We worked with Toledo to put together a map of the entire city with predictions, address by address online, so anyone can access the information at any time,” Schwartz says.
Related: Researchers Find Lead Exposure Linked to IQ Loss And according to Riley, the portal is getting a lot of attention thanks to media blasts and a YouTube video highlighting the research while helping customers identify pipes in their home.
It’s important to Riley that the public is involved, knowledgeable and feels comfortable asking questions when crews are working near their homes.
Keeping citizens feeling comfortable throughout the process of replacing lines is also a priority. “One of the key things we’re doing during the pipe replacement is providing residents a water pitcher with filter at no charge,” Riley says. “That way they know that while we are doing the work, they have safe water.”
The accuracy of the predictive models has also been a significant factor for securing funding.
“We have been able to leverage the BlueConduit work to mobilize more than $10 million in funding to replace lead pipes,” Riley says.
So not only is the software helping locate lead pipes, but it’s also allowing the city to act in a much faster way than originally anticipated. Changing the game
Before working with BlueConduit’s software, the city of Toledo predicted it would take around 30 years to replace all the lead service lines in Toledo, but now with the predictive model, they are hoping to achieve that goal in about seven years.
“Because we’ve been able to leverage the BlueConduit work to rally significant funding, our timeline for replacement has significantly advanced,” Riley says.
“Without BlueConduit, we would have had to rely on historical records and inefficient ways of checking pipe. This would have cost the city more money and likely placed a bigger burden on our ratepayers.”
Riley and the city of Toledo will continue to use the software until the goal of complete lead service line removal is achieved.
Source: BlueConduit company



