Improve miss_summary by providing a nice print method

[njtierney]

As discussed with @dicook, The aim with this is provide a ricer overall summary of missingness, providing information on:

• number of obs
• number of variables
• number of missing values
• prop missings
• number/prop of variables with missing values
• number/prop of cases with missings
• some 5 number summary of missingness

This could also be used iteratively so you can run this summary after making changes to missings and see what the impact is.

I think the way to do this is to write a print method for naniar::miss_summary(airquality), which is a function that does nearly every missingness summary. So you still get the underlying data when you do the summary, but it is presented in a really nice way.

[runkelcorey]

Workshopping this.

> naniar::miss_summary(airquality)
153 observations of 6 variables
44 (4.8%) of 918 values are
	* NA
	* -99
	* ""
2   of 6   (33.3%) variables have missing values
252 of 918 (27.5%) observations have missing values

For a 5-number summary, you could add a min/max/median amount of missing values for variables and for cases, but imho a compact print method is preferable. Many of us still work on our laptop, where screen real estate is a premium.

2   of 6   (33.3%) variables have missing values
Missing values are distributed throughout variables:
min median max
2   5      5