Addition of Geosilhouettes

[AnGWar26]

Thought it would be best if I gave you the opportunity to comment on this as I work @knaaptime.

Todo:

• [x] All this code is held together by running gdf.fillna(0) before running the clusterer, need to find solution that handles NaN values and keeps the index the same length as the dataframe.
• [x] Handle input when an aspatial model is passed to a spatial diagnostic
  • Currently throws what might be a cryptic error to most users 'NoneType' object has no attribute 'sparse', referring to the absence of a W object
• Figure out why these functions return 3 axes when I think they should be returning 2. The third is dtype=object

Updated 8/19

[codecov[bot]]

Codecov Report

Merging #256 into master will decrease coverage by 24.54%. The diff coverage is 13.54%.

@@             Coverage Diff             @@
##           master     #256       +/-   ##
===========================================
- Coverage   79.24%   54.70%   -24.55%     
===========================================
  Files          12       12               
  Lines         853      998      +145     
===========================================
- Hits          676      546      -130     
- Misses        177      452      +275     

Impacted Files	Coverage Δ
geosnap/_community.py	39.28% <5.69%> (-26.18%)	:arrow_down:
geosnap/analyze/analytics.py	65.75% <43.75%> (-6.43%)	:arrow_down:
geosnap/analyze/dynamics.py	22.58% <0.00%> (-77.42%)	:arrow_down:
geosnap/io/storage.py	24.66% <0.00%> (-70.67%)	:arrow_down:
geosnap/io/util.py	48.83% <0.00%> (-30.24%)	:arrow_down:
geosnap/_data.py	74.11% <0.00%> (+3.52%)	:arrow_up:

---

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 b1903b6...641fdb6. Read the comment docs.

[knaaptime]

cool, lets talk about this a bit during the meeting on tuesday. It'll be a little easier to describe what i mean about stashing the index so we dont have to fillna

Also i think it would be better to switch the ordering of the plotting nomenclature, e.g. instead of ***_plot(...), i think it would be better to do plot_***(...)

[knaaptime]

Handle input when an aspatial model is passed to a spatial diagnostic Currently throws what might be a cryptic error to most users 'NoneType' object has no attribute 'sparse', referring to the absence of a W object

one thing we could do is add a little metadata attribute to the ModelResults class that stores whether the model is spatial or aspatial. That way we can add a simple check if users pass the wrong type of model to the wrong diagnostic

[knaaptime]

awesome! looks like this has everything we talked about yesterday. Everything working as expected?

[AnGWar26]

As far as I can tell everything works as expected, and at this point I would designate this ready for review. However, this is such a big PR (+500 lines), and with your understanding of these statistics being better than mine, I would encourage you to test it out yourself before merging.

Let me know if these needs anymore changes.