Last month, the state of Michigan declared Flint’s drinking water quality "restored." To get to this point, it’s taken, among other things, more than 30,000 water tests.
Most of the data comes from the first half of 2016, just after the crisis became national news. Testing has slowed down significantly since then, but it’s still going: more than 350 samples have been tested this year so far.
When you remember that every data point represents somebody sending off a sample bottle of their water and waiting to hear back about whether or not it was safe, it really hammers home what the city went through.
When you look at the full range of measurements (top panel of the graphic), it's hard not to notice the sheer ridiculous amount of lead in some of the samples. Dozens have measured more than 1,000 parts per billion (ppb) - the highest was almost 23,000, from a sample taken in April of 2016. Those numbers are outliers: less than one percent of all the samples collected, but they're still shocking. While there’s no safe level of lead, the federal standard is just 15 parts per billion.
That said, most of the measurements (more than 90%) taken during this time period were less than 45 parts per billion, so the bottom panel of the graphic zooms in on that range to show more individual points and the overall progression of water quality.
From a regulatory standpoint, the water quality of a city is defined by its 90th percentile; that’s the lead level that 90% of samples test below for a given time period. The samples used to calculate this number, called "Tier 1" samples, come from the highest risk homes.
Things have definitely gotten better. For the first half of 2016, the 90th percentile was 20 ppb -- so it was above the federal standard. For the second half of 2017, it came in at six parts per billion. It’s been below the federal standard for 18 months now. So, when you hear the state saying Flint’s water quality has been “restored,” that’s what they mean.