After the jump are my projections for North Carolina extrapolated to a county-by-county level, along with some basic demographic information about each county. Since returns come in by county rather than Congressional District -- I figure it will be a little easier to follow the action this way. However, because my model was not designed to work at the county level, there may be an additional margin of error here. The results for the large counties should be more reliable than the results for the smaller ones.
Very important: These results are calibrated in order to match the current margin in the Real Clear Politics polling average: Obama +8.0. For Clinton to win North Carolina, she'll need to be beating these numbers by about 8 percent across the board. For Obama to win North Carolina by double digits, he'd need to better these numbers by a couple of points.
Also, an extra caveat when interpreting North Carolina returns. If the results of early balloting are released before most of the election-day vote is counted, the initial numbers might look artificially high for Obama.