search

arilamstein / censusdis-streamlit

App to explore the impact of Covid-19 on US county-level demographics.
https://census-explorer.streamlit.app/
MIT License
16 stars 7 forks source link

Any other variables that I should consider adding? #21

Open arilamstein opened 5 months ago

arilamstein commented 5 months ago

Ask around (ex Census Slack, other Census experts you know) to see if there are any other variables you should consider adding to the app. I only learned "WFH" (which is fascinating) was available because I saw someone else use it on linkedin for something else. There might be low hanging fruit here.

arilamstein commented 5 months ago

I just added in 'B19057_002E' : 'Total With Public Assistance'.

Unemployment seems interesting, but B23025_005E seems to only start in 2011.

Poverty seems interesting as well. I should probably ask people who are more knowledgable in ACS for ideas. It is now trivial to add another dataset: just add it to the variable census_vars and rerun gen_county_data.py.

vengroff commented 5 months ago

Looking good. It's nice to see how this project moved from a very basic beginning to where it is today.

One thing you could try, if you want to go this way, is to plot not just survey results directly from the U.S. Census, but functions of those. For example, the census does not compute diversity and integration metrics, but you can compute them, according to any one of several definitions, and plot the results. See https://github.com/censusdis/censusdis/blob/main/notebooks/Nationwide%20Diversity%20and%20Integration.ipynb for an example notebook or https://datapinions.com/diversity-and-integration-in-america-an-interactive-visualization/ for an interactive, yet static, web map. You could also plot other approaches to this concept, as described in https://datapinions.com/on-the-diversity-of-similarity-and-the-similarity-of-diversity/.

arilamstein commented 5 months ago

Looking good. It's nice to see how this project moved from a very basic beginning to where it is today.

One thing you could try, if you want to go this way, is to plot not just survey results directly from the U.S. Census, but functions of those. For example, the census does not compute diversity and integration metrics, but you can compute them, according to any one of several definitions, and plot the results. See https://github.com/censusdis/censusdis/blob/main/notebooks/Nationwide%20Diversity%20and%20Integration.ipynb for an example notebook or https://datapinions.com/diversity-and-integration-in-america-an-interactive-visualization/ for an interactive, yet static, web map. You could also plot other approaches to this concept, as described in https://datapinions.com/on-the-diversity-of-similarity-and-the-similarity-of-diversity/.

@vengroff I am very interested in going beyond just reporting the existing numbers. That's been on my wish list for a few years now. It might have to wait a bit though - I just submitted a proposal to talk about this project at an upcoming meetup. So right now I'm just trying to grab as much low-hanging fruit as possible!

vengroff commented 5 months ago

Whenever you are ready let me know. I'd be happy to talk you through the divintseg repo that does the computations I mentioned and more.