This is wrong on so many levels, from data through process to evaluation. How do you even prompt claude not to give you Pearson for correlating them.