ubicenter/blog

UBI Center blog, hosted via Jupyterbook.

ubi-center

Generates data for UBI Center's Plan Explorer tool and visualizations.

Installation

Install with:

pip install git+git://github.com/UBICenter/ubi-center.git

Questions

Contact the maintainer, Max Ghenis ([email protected]).

Citation

You may cite the source of your analysis as "ubicenter release #.#.#, author's calculations."

blog's Issues

Post on Black-white poverty and income gaps

For MLK Day or sometime in February (Black history month)

Similar to my June 2020 piece but with (a) new data, (b) plotly, (c) other metrics like poverty gap and mean/median income gaps

Change "Search this book" to "Search"

Black-white wealth gap post

For MLK Day or sometime in February (Black history month)

Similar to my June 2020 piece but with (a) new SCF data, (b) plotly, (c) potentially other wealth metrics (like K-S test's D statistic??)

See also #53 on the income side (I think income and wealth would be worth separating). cc @ngpsu22

Consider fastpages instead of jupyter-book

The jupyter-book lead developer mentioned in an issue that:

I think we have to walk a balance between re-creating all of the features of a blogging platform (since that's not technically in-scope of Jupyter Book) if it adds a lot of maintenance burden

I'm also concerned about how often jupyter-book requires rebuilding, at least when testing so far. This won't be sustainable with many blog posts.

This prompted me to look for alternatives that are geared more at blogging, but still with jupyter-book features like hiding/showing code, Jupyter input, interactive visualizations, etc. I found http://fastpages.fast.ai, which looks like a pretty great option. The only potentially missing feature is plotly support, and it also worries me a bit that there isn't much in the https://github.com/fastai/fastpages repo. But worth trying out I think.

Return figures rather than fig.show inside functions

To make work with plotly-converter @mgilbert1

Point users to ubicenter.org posts

Eventually we'll deprecate blog.ubicenter.org now that posts are on the main domain. Provide a warning and link in the meantime.

Change the bar colors of the PWD poverty chart

Ideally should be

• white background
• gray for people out of poverty (no legend needed)
• blues for poverty rates, a la the below chart

Restyle PWD benefit chart

• White background
• Blue for disability (maybe dark blue for the "any" program line)
• Gray for no disability
• Remove y axis title
• Change legend text
• Change axis to go to 100% (since it's close already)

Colab launch button doesn't work

It's pointing to https://colab.research.google.com/github/UBICenter/blog/blob/jb/20200707/adult_child_ubi.ipynb

where it has to point to https://colab.research.google.com/github/UBICenter/blog/blob/master/jb/20200707/adult_child_ubi.ipynb

(i.e. add the master branch to the URL github/UBICenter/blog/blob/master/jb/20200707/adult_child_ubi.ipynb)

Reverse order of PWD poverty bar chart

Start y axis at zero in charts in racial wealth gap post

@fedderw

Add link to UBI Center homepage

Until executablebooks/jupyter-book#740 is resolved and the logo link can go there

Limit taxes on AGI and taxable income to positive income in all posts

cc @ngpsu22 we should update old ones

Add Nate's blog post on child allowance and UBI

White background on Medium version of child/adult/all post

@ngpsu22 could you please edit the Medium post to use white-background charts with B in the x-axis from the notebook post? It's currently gray:

Veterans' Day post

@ngpsu22 is doing the coding, I'm doing the writing. Nate, could you share your notebook?

Fix Python style issues in adult_child_ubi.ipynb

As listed by the bot in #9 (comment)

Add Google Analytics tracking

See https://jupyterbook.org/customize/config.html?highlight=tracking#configuration-reference

Add GitHub and launch buttons

Update plotly in blog posts for Colab

Running https://colab.research.google.com/github/UBICenter/blog/blob/master/jb/20200707/adult_child_ubi.ipynb shows this error. I think it's because it's running an old version of plotly, so Colab notebooks should update plotly. cc @ngpsu22

ValueError                                Traceback (most recent call last)
<ipython-input-1-c893d11fa016> in <module>()
    201            title='Overall poverty rate and spending on cash transfer programs',
    202            xaxis_title='Spending in billions',
--> 203            yaxis_title='SPM poverty rate')

4 frames
/usr/local/lib/python3.6/dist-packages/plotly/basedatatypes.py in _raise_on_invalid_property_error(self, *args)
   3664                     full_obj_name=full_obj_name,
   3665                     invalid_str=invalid_str,
-> 3666                     prop_descriptions=self._prop_descriptions,
   3667                 )
   3668             )

ValueError: Invalid property specified for object of type plotly.graph_objs.Layout: 'legend_title_text'

Remove ipynb_checkpoints

And add as a .gitignore