New Report Presents Method for Data Driven Prediction of Lead Service Lines

A recent report by University of Michigan researchers outlines a modeling approach for replacement of lead service lines that was developed from their participation in the Lead Service Line replacement program in Flint Michigan.  This approach can help to target the replacement of lead service lines, maximizing the use of available funding.  Using property data, historical data, lead test results, and data from physical inspections, the modeling estimates the private and public side service line materials.  An “active learning” phase generates a set of homes for actual inspection.  Finally, algorithms determine which locations should be targeted for lead service line replacement.  The authors believe this process could be used in other locations that want to make the most efficient use of funds in their lead service line replacement program.