We construct a public atlas of children's outcomes in adulthood by the Census tract in which they grew up using anonymized longitudinal data covering nearly the entire U.S. population. For each tract, we estimate children's earnings distributions, incarceration rates, and other outcomes in adulthood by parental income, race, and gender. Children's outcomes vary sharply across nearby areas: for children with parents at the 25th percentile of the national income distribution, the standard deviation of mean household income at age 35 is $5,000 across tracts within counties. We illustrate how the tract-level data provide insight into how neighborhoods shape the development of human capital and support local economic policy using two applications. First, we show how these data can be used to better target policies to improve economic opportunity by uncovering specific neighborhoods where certain subgroups of children grow up to have poor outcomes. Neighborhoods matter at a very granular level: conditional on characteristics such as poverty rates in a child's own Census tract, characteristics of tracts that are one mile away have little predictive power for a child's outcomes. Second, we show that the observational estimates are highly predictive of neighborhoods' causal effects, based on a comparison to data from the Moving to Opportunity experiment and a quasi-experimental research design analyzing movers' outcomes. We then identify high-opportunity neighborhoods that are affordable to low-income families, information that can be used to design affordable housing policies. Our measures of children's long-term outcomes are only weakly correlated with traditional proxies for local economic success such as rates of job growth, showing that the conditions that create greater upward mobility are not necessarily the same as those that lead to productive labor markets.