DOC: Fixed some pep8 violations, arr.size instead of len(arr)

[tommyod]

I read through some files in statsmodels/nonparametric and found some pep8 violations which I fixed. Very small stufff, mostly spaces. Remoted a few unused imports. Made one change to the code: using arr.size is faster than len(arr).

Comments very welcome.

[coveralls]

Coverage increased (+0.0006%) to 82.171% when pulling 69b05ef on tommyod:kde_review into 0d757e5 on statsmodels:master.

[codecov-io]

Codecov Report

Merging #4219 into master will increase coverage by <.01%.
The diff coverage is 84.84%.

@@            Coverage Diff             @@
##           master    #4219      +/-   ##
==========================================
+ Coverage   79.56%   79.56%   +<.01%     
==========================================
  Files         562      562              
  Lines       83843    83840       -3     
  Branches     9553     9553              
==========================================
- Hits        66711    66709       -2     
+ Misses      14948    14946       -2     
- Partials     2184     2185       +1

Impacted Files	Coverage Δ
statsmodels/nonparametric/bandwidths.py	92.85% <ø> (ø)	⬆️
statsmodels/nonparametric/kdetools.py	70.96% <83.33%> (ø)	⬆️
statsmodels/nonparametric/kde.py	85.81% <85.71%> (+0.39%)	⬆️
statsmodels/stats/descriptivestats.py	24.13% <0%> (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 0d757e5...69b05ef. Read the comment docs.

[josef-pkt]

Made one change to the code: using arr.size is faster than len(arr).

Is there much of a speed difference? How does shape[0] compare?
The problem is that size has a different meaning than len or shape[0]. Even if it's not currently used, I prefer code that would allow extensions. (Less to think about when code is changed because our standard is to have observations in rows.)

And thanks for pointing out the non-style changes in the description. That makes it much easier to quickly check the changes.

[tommyod]

@josef-pkt Upon further investigation, the speed difference is practically insignificant. I have removed this code change from the PR. The unused imports and the pep8 improvements are still present.

[josef-pkt]

@tommyod Thanks
Can you squash this into one commit, then it looks ready to merge.

pure pep-8 style changes, commit prefix STY
we roughly follow this
http://www.statsmodels.org/stable/dev/maintainer_notes.html?highlight=prefix#commit-comments

[josef-pkt]

merged, Thanks @tommyod