Suggestion: move "observation" legend at top of theoretical distribution legend

[gmgeorg]

I have a suggestion for the "Cullen and Frey graph" to be a bit less confusing: the legend on the right should include also the "[blue dot] observations" marker ( on top of or below all the theoretical distribution).

Right now, at first glance it appears that the observations are falling at the top left corner of the graph with a skewness^2 of 1 and a kurtosis of ~1.2. It took me multiple checks to realize that that is the legend, not the actual data point in the (kurtosis, skew^2) plane. This is further exasperated if data follows closely to a normal distribution as then the actual blue dot kind of disappears behind the Normal dot/star shaped pattern - and the only blue dot on graph is the legend ( but it's not clear at all it's a legend)

See e.g. https://stats.stackexchange.com/a/629866/11476

I ll attach an example adaptation of the plot that I think would be more clear to read / interpret.

[gmgeorg]

See here for an example of before/after, with some other nit improvements for visual clarity/increase data-ink ratio.

Before

After

and w/ boot = 100 for a heavy-tailed Lambert W x Normal distribution using

library(LambertW)
set.seed(10)
x <- rLambertW(n=1000, "normal", theta=list(beta=c(mu = 1, sigma=2), 
                                            delta=0.15))

[aursiber]

Thanks for the suggestion.
The legend is now improved (see d6eeb4c and f5a892b).