QIIME2 does not have a pearson/spearman correlation method for correlating features with metadata — the reason being that this is not really an appropriate test for compositional (e.g., relative abundance) data.
You may want to check out q2-gneiss. It is a complicated but powerful method that could be used to approach this question (correlate metadata with all features).
When you have your post-intervention data, check out q2-longitudinal — some of the tools in there are designed for pre/post studies.
I hope that helps!