update nbsit notebook

[renanxcortes]

Cool! This notebook is a huge improvement! I'm trying to reproduce the notebook here and am getting this from cenpy (do you know what might be happening?):

[renanxcortes]

I've seen that products is a branch of cenpy, right?

[knaaptime]

i added a note. you can either use the github version of cenpy and checkout the products branch, or you can test out the first alpha with pip install --pre cenpy

[renanxcortes]

Ok, cool, I managed to install the pre-version, but when I run the third cell I get:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-5-b813ef537da5> in <module>()
----> 1 atl = products.ACS(2017).from_msa(name='Atlanta, GA', variables=variables.copy(), level='tract')

~\AppData\Local\Continuum\anaconda3\lib\site-packages\cenpy\products.py in from_msa(self, name, variables, level, **kwargs)
    357     def from_msa(self, name, variables=None, level='tract', **kwargs):
    358         return self._from_name(name, variables, level,
--> 359                                'Metropolitan Statistical Area', '_msas', **kwargs)
    360     from_msa.__doc__ = _Product.from_place.__doc__.replace('place', 'MSA')
    361     def from_csa(self, name, variables=None, level='tract', **kwargs):

~\AppData\Local\Continuum\anaconda3\lib\site-packages\cenpy\products.py in _from_name(self, place, variables, level, layername, cache_name, strict_within, return_bounds, geometry_precision)
    326                    strict_within=True,
    327                    return_bounds=False, geometry_precision=2):
--> 328         if level not in self._layer_lookup().keys():
    329             raise NotImplementedError('Only levels {} are supported. You provided {}.'
    330                                       'Try picking the state containing that level,'

TypeError: 'dict' object is not callable

[renanxcortes]

Trying to debug here.. this must have something to do with from_msa from these functions: https://github.com/ljwolf/cenpy/blob/product/cenpy/products.py

[knaaptime]

that's been fixed upstream if you use the github version

[renanxcortes]

hm... so I reinstalled cenpy running pip install --pre cenpy and then from cenpy import products worked smoothly... what would be the github version in this case?

[knaaptime]

the products branch has an early RC available on pip using the --pre flag, but that release has a small bug in the _from_msa method. That bug is fixed upstream in @ljwolf's repo, so to use it you

1. uninstall any cenpy's you currently have
2. clone https://github.com/ljwolf/cenpy.git
3. git checkout products
4. python setup.py develop (note that since you're using the develop command, the version you import will change depending on which branch you have checked out)

[renanxcortes]

the products branch has an early RC available on pip using the --pre flag, but that release has a small bug in the _from_msa method. That bug is fixed upstream in @ljwolf's repo, so to use it you

1. uninstall any cenpy's you currently have
2. clone https://github.com/ljwolf/cenpy.git
3. git checkout products
4. python setup.py develop (note that since you're using the develop command, the version you import will change depending on which branch you have checked out)

Ow, I see, thanks! I'll try here!

[renanxcortes]

It worked! Thank you! :) I'm gonna try to reproduce now! (I'm downloading the network...)

[knaaptime]

were you able to reproduce?

[renanxcortes]

So, I still didn't reproduce the notebook on my side (I'll keep trying), but I think this is good to be merged will some small tweaks:

• update comment of cenpy (perhaps also adding the python setup.py develop part) to be before call on the first cell

• Is there a way to add an external link for the source for the list of the variables names in the notebook? For example, if the user wanted to change the group (other than black, white, hispanic, asian) where could he find the respective codes?

• add an additional comment that calc.access could take some time to run (similar to the comment on downloading network)

• change "balt" on :

# balt_network.save_hdf5('atl_network.h5')

[knaaptime]

excellent, thanks

[renanxcortes]

No sweat! Ow, also consider using the recent Multi_Information_Theory class for the last function. I wanted to check if the values were matching.

EDIT: I checked here and, indeed, the values matched:

[renanxcortes]

I think I managed to run the notebook reducing the buffer to 1000. Since this is a tutorial notebook for explaining functions, could we change this parameter?

[renanxcortes]

Also, shouldn't nbsit function should be run in atl_access instead of atl? I think with the latter it is building the aspatial version of the index, isn't it?

[knaaptime]

lol yeah, oops. I had both in there for comparison (the stats are quite different) but apparently deleted the wrong cell when I tried to simplify. Good catch

[renanxcortes]

Also, geopandas is being imported two times. I think you can drop one.

[renanxcortes]

Awesome, thanks for adding these new functionalities, @knaaptime !