Friday, February 26, 2016

The Carnival Weight Guesser







I'm a big fan of "correlation does not imply causation".



And, sometimes, it's fun just to wallow in the statistic no matter how meaningless they might be. This YourMechanic article has lots of little tidbits of love in it for those willing to get dirty. Hey, that's me.


So, mostly, the point is the headline, "Do Republicans and Democrats Drive Different Types of Cars?" which is cute but the actual numbers devolve into a red state blue state compare and contrast which immediately adds a second level of abstraction with the "data". And the car numbers are already a bit skewed because they come from one source, "YourMechanic".


But let's have some fun anyway...
So this is exactly what it looks like. These are the most popular vehicles, based on numbers from the cited source, in red and blue districts. Wow, you say, the lists are almost identical and yes, yes they are. It's not what you thought it would be is it? Well, it's not until we adjust the conditions a little bit. If we take the vehicles that are unusually popular in each district, remove the national big sellers and see if that tells us something different?
And Bingo! We get results that fit our expectations, which isn't the most accurate way to derive information from data but, well, it's fun.

And then there's this...

Which, again, sort of dovetails in with what we expect. I want to add, one final time, this sort of statistical manipulation isn't nearly as meaningful as some people might want but it is, once in awhile, kind of fun. Have a fun Friday out and drive safe.