Did some research on this because im a political news junkie on top of being a massive nerd with some degree of understanding of math.
If you go back to the 2016 election and average out the last 5 polls of the state that weighed Trump vs Hillary, the average poll indicated that Hillary would beat out Trump in the state by an average of about 6.2% (31 points in favor of Hillary / 5 polls total = 6.2 in Hillary's favor. Last 5 polls only one used since all other polls happened almost 6 months before the election even happened)
So assuming the bias in polling stays about the same (its an inexact science to begin with that is aggregated among many different poll takers with different procedures), we can next argue that if Trump is within about 5% to 10% of the potential Dem nominee, it makes it at least possible for Trump to win the state..... A 5% difference between two candidates puts the state well into the 'swing state' category where it could go either way, and and if we take into account the possible poll bias that favors Dems in the state, even a 10 point deficit is not a deal-breaker for Trump.
Now we just use state polls taken for the state that specifically look at matchups between Trump and individual candidates. There are 4 that show up:
Assuming that Klobuchar doesnt become the nominee, and that Biden, Warren, and Sanders will likely be the nominee, then Trump still has a pretty uphill battle to climb.
If we apply the polling bias to the leads that would swing things back in Trump's favor by 4.7%, that puts him down 8% to Biden, down 7% to Warren, and down 4% to Sanders.
The huge ass 'IF' here that hangs over the logic of this whole theory is that the current matchups between Trump and the Dem candidates will inevitably change. Only if the election were held RIGHT NOW would Biden and Warren likely win the state while Sanders has to fight for it a little, that could certainly change by election time, though in whose favor is anyones guess given how we dont even have a nominee yet and that many more scandals and policy shifts can take place between now and the election that is next year.
To reiterate: Trump faces a big uphill battle, but its not out of the question that he could win the state depending on the nominee and what happens between now and election day. The GOP winning Minnesota is certainly not out of the question compared to some other states that Trump and the GOP would try to flip. (Virginia for example was actually one of the most accurately polled states in the race, and it was a full 5 points in favor of Hillary when almost every other swing state was won by Trump.
https://www.realclearpolitics.com/epolls/2016/president/va/virginia_trump_vs_clinton-5542.html )