1) I think Trump will win North Carolina based on the trend set in 2012, the state itself, and the amount of money and time that the RNC has invested there. It's possible I am wrong though because 2008 was a shocker.
1a) I also think Trump will win in Ohio (he has had a consistent and positive set of polls there), a win in Iowa (again consistent and positive polls), and he has to win FL. He may win FL by sheer virtue of Marco Rubio's strong Senate campaign though.
2) Recent polling (including my undergraduate school CNU) has suggested a Clinton +12-+15 lead in VA. I actually called up local talk radio to discuss this the other day.
It's not going to be that big. here is why: http://fivethirtyeight.com/features/mark-warner-why-polls-missed-a-shocker-in-virginias-senate-race/
VA polling is usually extremely skewed because of Northern Virginia. It is highly likely that Clinton will win VA, but it won't be by 15 points. More like 1-5 points, but we will see.
Arizona is surprising. But, some recent polls out show anywhere between a Clinton +5, to a tie, to a Trump lead. It's all over the place. Reliable and respected polls have not be conducted there. If you look at 538 some of the polls are graded as C-.
Wisconsin, I can see a Clinton +7 if it is a blowout. If it is a close election it may be more like 2-3 Clinton. I don't think Trump wins Wisconsin.
New York, yeah. I am not surprised.
Utah, if you look at 548 you will see it's all over the place. http://projects.fivethirtyeight.com/2016-election-forecast/utah/
One poll has McMullin up by 4 and another has Trump up by 15. The trend is still Trump in Utah.
+5 Clinton in Colorado is similar to Obama tier levels from 2012. It wouldn't surprise me if that was the case.
Most recent poll from Nevada shows Clinton +4, but another shows Clinton +7. Remember, Obama won the state twice. However, long term polling has shown Trump leads in Nevada, so we will have to see if any of that returns.