Motivation
I am interested on using R to do data analysis on a considerable amount of data. Having to plot each time I want to observe if it fits any particular distribution by eyeball (which can be misleading), and recall in memory how the shape looks like (or compare, say, on a book).
As a programmer, I would like to be able to have the program instead looking this to me. Standardized moment was something I studied in an undergraduate course and seems that Kurtosis is of particular interest of the community according to my search.
Questions
Wikipedia -- Kurtosis seems to have a considerable value on them that is very precise on saying the distribution simple based on the number. How far can I trust this instead of eyeball the plotted data? If this is not a yes/no answer, how can I claim that after a given value it is acceptable to apply, say for example Linear Regression assuming Kurtosis indicated a normal distribution (given the other constraints are correct)?
The same as above, but for moment 3, Skewness. I would also appreciate if someone would point how to understand if it is left or right skewed or symmetric based on the values of 'the method'.
When I mention 'the method' above I noticed wikipedia suggest few of them. So the third question is, is there currently a chosen one or depends on the context? If depends on the context, how can I go about analyzing?
I would appreciate if someone could point an R function/package that does this as well.
Thank you.