Deal with NaN values in LatLong columns

[thehomebrewnerd]

Closes #453

This PR serves as replacement for #693

If a single Nan value was present instead of a tuple in a LatLong columns, this could result in indexing errors when using the Haversine, Latitude or Longitude primitives. This PR fixes that issue.

A check was put in place during entity creation to replace any NaN values with (NaN, NaN). This check is also implemented in the Haversine, Latitude and Longitude primitives in case the primitive functions is called directly as in the doctests.

[codecov]

Codecov Report

Merging #1007 into master will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1007      +/-   ##
==========================================
+ Coverage   98.28%   98.29%   +0.01%     
==========================================
  Files         121      121              
  Lines       12397    12474      +77     
==========================================
+ Hits        12184    12261      +77     
  Misses        213      213              

Impacted Files	Coverage Δ
...retools/primitives/standard/transform_primitive.py	100.00% <100.00%> (ø)
featuretools/tests/entityset_tests/test_entity.py	100.00% <100.00%> (ø)
...s/tests/primitive_tests/test_transform_features.py	98.80% <100.00%> (+0.08%)	⬆️
...eaturetools/tests/utils_tests/test_entity_utils.py	100.00% <100.00%> (ø)
featuretools/utils/entity_utils.py	98.29% <100.00%> (+0.09%)	⬆️

---

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 0f35678...9f4c1b9. Read the comment docs.

[jeff-hernandez]

Looks good!

[rwedge]

Looks good!