If the trend shown is very close to what the national trend is (which I suspect it likely is), then that is the trend you would expect to see. For instance, if you were charting rates of deaths from illnesses by county, you would expect all of the counties to have increased rates for elderly people, because it is a fact that old people die from illness more often. You would not expect to find one or two random counties where the elderly people in that county have significantly lower rates of death, because that would be against the average. If you did find an anomaly like that, either you have found an oasis where old people live forever, or you found data that is probably inaccurate. So if the voting pattern that you graphed is consistent across the country, the interesting (and possibly false) data would be the anomalies that stick out as unusual. For instance, if you were to chart Georgia or Michigan, my guess is you would find a couple of counties (I bet you can guess which ones) that have a different trend line, which would indicate cheating (the counties that had over 100% voter turnout, for example).