The explore module computes statistic measures for correlation, but some of them have assumptions, e.g.
PCC assumes that both variables are normally distributed.
Add a check and a warning if requirements are not met.
Instead of warning, adding distribution/statistic checks into reports/results would be useful too. It also supports bias checking (if the data is not normally distributed or skewed in particular characteristics).
The explore module computes statistic measures for correlation, but some of them have assumptions, e.g. PCC assumes that both variables are normally distributed. Add a check and a warning if requirements are not met.