Here's their methodology:
we used the White House’s “Job Impact by Congressional District” (which has had its share of criticism for overestimating the new job totals) to determine the total number of new jobs created by state, and then compared those numbers to the total unemployed persons in each state. The result is a hypothetical percentage of each state’s unemployment that will be solved by the stimulus plan, and while of course this number will fluctuate based on upcoming layoffs, it at least gives a picture of how the proposed job creations impact each state’s current unemployment problem.So the map doesn't depict the reduction in unemployment in absolute terms, but in terms of a percentage of the unemployment rate. That's why states like North Dakota and West Virginia come out looking like big winners, and places like California and the Carolinas not so much: as you'll recall from these maps, the unemployment rate is just much lower in the Plains than in the Southeastern or West Coast states, so the creation of x number of jobs per capita will reduce a greater percentage of those states' unemployed.
An ideal stimulus would cause this whole map to be shaded the same color, or would even disproportionately benefit places with higher unemployment. But as a practical matter, not to mention a political one, it's probably difficult to engage in very precise targeting of the most economically damaged parts of the country. Nonetheless, if these numbers are right, the stimulus really will have gone a long way towards reducing human suffering: no state's unemployment rate is reduced by less than a quarter or so, which means an awful lot of people will avoid perconal catastrophe. (Another debate could be had, of course, on the bill's long-term effects.)
Meanwhile, the gummint has a clickable map that will send you to state websites that describe how stimulus fundds are being spent in each state. Except for North Dakota, where the government does not care about its people.