The first lead pipe is replaced in Flint. Image: Bill Pugliano/Getty Images

In March, Flint, Michigan began the seemingly impossible task of replacing the pipes that delivered lead-poisoned water to its citizens. I say “impossible” because there’s no real way for the city to know with absolute certainty which pipes need to go. A team of data scientists from Google and the University of Michigan’s Michigan Data Science Team has developed an algorithm it claims can help.

Advertisement

The problem in Flint is the same issue that many older cities face when updating their infrastructure: There aren’t accurate records for when many service lines were installed, or even where specifically they are located. According to the University of Michigan, there are about 8,000 lead service lines in Flint but the city only knows the status of 4,376 of them.

When Google partnered with the school earlier this year to help develop any digital tools that might help in the crisis, it was apparent that focusing on repair efforts would be key. According to Jacob Abernethy, a computer science and engineering professor at University of Michigan, there had been plenty of data collected throughout the water testing process, but no effort to apply that data towards actionable real-world solutions.

The year in which the property was built is one of the most important indicators for lead. Image: Michigan Data Science Team

Using building records as well as the data collected by water testers, scientists created a predictive algorithm that makes a good guess about the state of each of Flint’s service lines based on certain property indicators, like the age, size, and type of building.

Advertisement

With the information they do know, the scientists will be able to deploy this algorithm across the entire city using machine learning, creating a fairly accurate map which recommends how repairs should be prioritized based on the number of people they will benefit. The team reached out to me today and passed along this detailed document with a summary of their findings for those who want to know more.

An interactive map shows the known status of all the city’s pipes and can guess the status of ones without data. Image: Michigan Data Science Team

In addition to feeding this information to city workers who are strategizing which infrastructure needs to be overhauled, the team has a secondary goal to deliver this information to the public. An Android app that will allow residents to have this data in-hand will launch this summer, and it will include additional information about lead testing results and recommendations for water filters.

Advertisement

Sponsored

[University of Michigan]