IDEA: take the spurious bivariate correlation website and run triplets through...

@mistersql

IDEA: take the spurious bivariate correlation website and run triplets through PyWhy and see if all the causal relations are 0.

I'm not smart enough to know if you can do causal analysis with bivariate data.

pywhy.org/dowhy/v0.11.1/exampl

Self-replies

Banging the drum that correlation isn't causation make statisticians think they are doing science education to discourage bad choice of model, when really, the effect is to imply that they are debunking science. "Meat, cigarettes and alcohol are all fine, keep consuming! Statistics is just arrogant smarty pants who think that correlation is causation and all science is cultural-warfare bunk!"

And this idea that statistics is incapable of discovering causal models has been going on for decades, like someone really, really wants people to think that causal relationships are impossible to discern, it is worth something to someone. No one gives a $#!+ if you misunderstand quantum chromodynamics, which you probably do.

"Correlation isn't causation" - True, but people want you to do things not in your personal best interest with this fact. We want you to only know this fact and become a chump.

Closely related to "True, but useless", which seems to come up at the office. "How should we solve this problem? Well it depends" True, but people who say that ain't following up with "on what" or they could have left out the prologue.