Drop missing data centrally while allowing marks to disable

[mwaskom]

I'm not totally certain what the best approach to missing data is; this will drop missing data after stats and scaling. I don't think seaborn needs explicit functionality for considering "missing" a valid category (it is just as easy to do df.fillna("missing")) but there may be times when it is desired to have missing data represented with (invisible) artists.

For now, that's not externally parametrize, but it is possible for a specific Mark to implement that behavior.

[codecov]

Codecov Report

Merging #2816 (693aa4b) into master (34b8873) will decrease coverage by 0.00%.
The diff coverage is 85.71%.

@@            Coverage Diff             @@
##           master    #2816      +/-   ##
==========================================
- Coverage   97.96%   97.96%   -0.01%     
==========================================
  Files          67       67              
  Lines       21756    21772      +16     
==========================================
+ Hits        21313    21328      +15     
- Misses        443      444       +1     

Impacted Files	Coverage Δ
seaborn/_marks/basic.py	60.86% <25.00%> (-1.04%)	⬇️
seaborn/_core/plot.py	95.11% <100.00%> (+0.02%)	⬆️
seaborn/_core/properties.py	99.15% <100.00%> (+<0.01%)	⬆️
seaborn/_marks/scatter.py	100.00% <100.00%> (ø)
seaborn/tests/_marks/test_scatter.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 34b8873...693aa4b. Read the comment docs.