"Count/proportion of missing values" per specified variable needs to be somewhere in toolbox (maybe summary statistics); and/or "filter out missing values"/"select only missing values" should be options in the columns

[gweinberg]

Problem

On https://isle.stat.cmu.edu/data-explorers/asteroids/, I'd like to see what proportion of the asteroids have names. Currently, there seems to be no simple way in the toolbox to get such info.

One way would be if summary statistics had a section for "count/proportion of missing values" and then the user could specify a variable.
It would also be nice if such functionality could be extended, to find "what count/proportion of variable _________ AND variable __________ are missing from data",
"what count/proportion of variable ________ AND variable __________ AND variable __________ are missing from data", etc.

Another way (maybe even better) would be if, in the top of columns, it was possible to select and/or filter out "missing". The filtering out would need to be explicit, like, when you click in select window at the top of a categorical variable, there should be an italicized option that says "filter out missing values"; and for quantitative columns, there should be a similar extra functionality at the top of the column, where you could click to "filter out missing values"

In fact, thinking a bit more, "select only missing values" would also be a useful option in columns, because, for instance in the asteroids data, I'd be interested to know whether the un-named asteroids differ meaningfully from the named asteroids. Are they smaller? Are they darker? I'd like to be able to first filter out the un-named and do descriptive stats, then go back and select only the un-named, and do the stats again. Currently, the only way to do this would be to use transformations -> combine categories -> create a new binary variable called, for instance, "named," and then, by hand, enter "yes" for every single name, and "no" for every single un-named, which is impossible since the dataset has 131,000 asteroids.

Alternatively, maybe the combine categories transformer should be updated so that it would be easier to select a whole series of many categories (e.g, they could appear in a window and you could select one, scroll to end, then shift-click to select all the categories within the range, like in Windows file-managers). Honestly, the combine/rename categories widget really needs to be redesigned anyway, it's rather confusing.

Solution

see above

Alternatives

No response

Additional context

None.