There is always some statistical noise in polls. When a new poll comes out with a huge swing from the last poll and nothing happened between the two that would cause a dramatic shift in public opinion, I tend to be skeptical of the new result, the previous result, or both. If it seems too good to be true it probably is.
On the whole, the bulk of the recent Wisconsin polling has shown Walker ahead. Polls have been wrong before and they can again, but I do think Walker will end up winning because the bulk of the polling has shown him ahead. I do not think he will win by 7 but again stranger things have happened.
Personally, I think that Obama's visits have hurt the candidates that been the presumed benefactors of those visits. It extends to Biden as well. When he visits, the intended benefactor doesn't receive any benefit from it and it actually hurts the candidate they support.
In NY-24, Dan Maffei went from a 8-point advantage over John Katko to a 10-point deficit in the poll released immediately after Biden's visit.
In the Wisconsin governor's contest, Obama visits to support Mary Burke and immediately afterward the polling shows Walker with a large spike in support.
Earlier, it happened in Maryland. Obama had gone to lend support to Anthony Brown. Immediately after the visit, Larry Hogan had a surge in the polls. Hillary Clinton was just there in support of Brown and I expect that Hogan could see another bounce in the polls.