Elio Campitelli

So... what distribution is this histogram?

#RStats

Jan 20 at 11:49pmPhanpy
Show threadElio CampitelliJan 21

More than one reply suggesting Laplace distribution. I digitised the church's outline and then tried to fit a Laplace distribution. Here's the fit.

#RStast #Iceland

Show threadElio CampitelliJan 21

Cauchy might be a better fit.

#RStats

Show threadWeekend EditorJan 21

@eliocamp

The main motivation to suggest Laplace, for me, was the pointy bit at the top. The tails are definitely too thick.

If we ignore the pointy peak, have you tried a t distribution, varying the number of degrees of freedom?

Show threadElio CampitelliJan 21
@weekend_editor Here's with t distribution added. Not super great!
Show threadEric BookJan 20
@eliocamp Laplace
Show threadVirginicusJan 21
@erc_bk @eliocamp Mais la Place est juste devant.
Show threadWeekend EditorJan 21

@eliocamp

Tails are really fat, judging by the heights at the edges. That would incline me to guess a Cauchy distribution.

However, the central part appears to come to a point? That would incline me to a Laplace distribution.

No idea how to reconcile those, other than "Architects don't care about building resemblance to distributions." :-)

Show threadElio CampitelliJan 21
@weekend_editor We should come up with the Hallgrímskirkja distribution.
Show threadKrzysztof SakrejdaJan 21
@eliocamp clearly a truncated cauchy
Show threadAngel FelizJan 21
@eliocamp I will assume normality 😁
Show threadAngel FelizJan 21
@eliocamp Laplace since to be the best fit
Show threadElio CampitelliJan 21
@AngelFelizF This is the best fit I could get (in terms of RMSE).