This script performs pariwise comparisons between variables in a data frame and saves some plots for the most extreme relationships with regard to a metric, e.g. the mean squared difference.
It's by no means formal hypothesis testing, but should give some intuition of how the variables in the data frame are related to each other.
$ python3 regtest.py -h
- replace this permutation ridiculousness with an actual distribution.
- implement a solver based on polyfit and classify relationship as linear if if best fit is with d=1