Good work Patrick!
I'm assuming that goals per game is all goals in a game not just those scored by the Nix?
Nah, it just the goals per game scored by the Phoenix. It's the Goals For column in the table for that season divided by the number of games.
It would be quite interesting to see if adding in Goals Against makes the correlation stronger or weaker... hmm.
EDIT: It makes it weaker, much much weaker (-0.1749).
However, if I use goal difference, it has a very very strong correlation (0.9541). When compared to the relationship between winning games and the attendance, this strong relationship suggests that if we thump teams in a few games (a la 6-0 vs GCU), the crowds will come, even if we have lost quite a few games by a small margin (i.e. 1-0, 2-1)