Does the map actually work? We checked it against real blood tests.

Our national map predicts childhood lead risk from nothing but public housing age and poverty data. To find out whether that prediction is real, we held it against states' actual measured childhood blood-lead, neighborhood by neighborhood. Across eight states it tracks, at the same strength the federal EPA study reported. Every number below is checkable. Click any state to see the two maps side by side.

0.77
rank corr
New Jersey
21 countys · strongest
→
0.70
rank corr
Wisconsin
208 tracts · Milwaukee, fully automatic
→
0.68
rank corr
Illinois
56 countys · Chicago + downstate
→
0.66
rank corr
Michigan
2,156 tracts · 4.2M-test research data
→
0.65
rank corr
Ohio
2,534 tracts · 4.2M-test research data
→
0.56
rank corr
Missouri
40 countys · St. Louis lead belt
→
0.55
rank corr
Iowa
56 countys · old rural housing
→
0.48
rank corr
New York
61 countys · NYC + upstate
→

How to check our work in five minutes

1
Open any state above. Left map is what we predicted from housing and poverty. Right map is the real measured blood-lead rate. Same geography.
2
Look at the dark zones. Where our prediction is high and the measured rate is also high, the map worked. In New Jersey, Wisconsin, and Illinois the cores line up tightly.
3
Check the number. The rank correlation is how closely the two maps agree, from 0 (random) to 1 (perfect). The federal EPA validation landed at 0.49 to 0.63. We land at 0.48 to 0.77.
4
Trace the sources. Measured data is each state's own surveillance, pulled from the CDC tracking network or the state's open API (Wisconsin, New York). The method is the EPA-published housing-plus-poverty approach.
5
Note what is honest about it. This predicts where risk concentrates, not which child is poisoned. It is a screening tool. That is exactly why it ends at an actual at-home test.

The method, in one paragraph

Old housing carries lead paint and lead pipes. Poverty tracks deferred maintenance and older stock. We pull both from the U.S. Census for every county and census tract, weight the oldest housing the way the Washington State Department of Health does, and score each place against the rest of the country. A 2024 EPA study validated this exact approach against roughly 4 million children's measured blood-lead levels. Our independent eight-state check confirms it holds. The full methodology and citations are in the white paper.

Method anchor: Zartarian et al., "A U.S. Lead Exposure Hotspots Analysis," Environmental Science & Technology, 2024 (DOI 10.1021/acs.est.3c07881). Measured data: CDC National Environmental Public Health Tracking Network; Wisconsin DHS and New York DOH open data; Michigan and Ohio tract data via the EPA study's Supplement B. Predicted risk: U.S. Census ACS 2022, scored by DetectLead.

The map finds the risk. The test confirms it.

See where your own neighborhood lands, then check it for real.

Open the national map Get the test kit