Identify meaningful demographics to enhance "tree equity"

[allenbaron]

Demographics info is available down to the census block level.

Exploratory analysis is needed to identify demographics could appropriately be used to enhance "tree equity"

[icanhazcodeplz]

data/demographics_block_group.csv provides race and in_poverty data at the census block group level. Age breakdowns could easily be added (look at scripts/create_demographics). Unfortunately, it does not look like ACS collects language data at either the tract level or block group level.

I think we should ask people at EcoAction which metrics they think would be valuable before spending too much time on this. They might be happy with just pct_nonwhite and pct_in_poverty, which we already have.

[Chellison]

I can make a start on this and share progress at the next meeting.

[Chellison]

Made a start for displaying a few different demographic and canopy cover variables across census areas (block group and tract level). Code and maps are in a pull request. Maps and a few graphs are on the google drive (data -> demographics_canopy_variables_per_censusArea).

The main remaining questions/to-dos are:

• Total population of different categories aren't the same? which to use?
• Remove census areas that cover Arlington cemetery and Reagan National?
• Get neighborhood names for tract instead of geo_id
• make code more generalizable and less hard-coded if necessary
• add canopy coverage as layer (got error when tried to, maybe too large)

[allenbaron]

EcoAction specifically mentioned using block groups with low income in a recent email:

I envision identifying the blocks with the least amount of trees and lowest income levels and assigning people to pound the pavement in those groups

They also wanted income and race data at the civic association level.

[allenbaron]

@Chellison, Tree canopy coverage might not plot correctly if it has coordinates in a third dimension (i.e. Z = elevation). Those aren't important and can be removed by running st_zm() without any options. I discovered this when trying to map some of the original shapefiles that were placed in Google Drive.

It is also a VERY detailed layer, which makes it challenging to include. I had to download QGIS to see it mapped. I never got R to work.

[Chellison]

Thanks for the tip, I'll look into that function!

I'm in the same boat with getting it to map with other software but not in R/Leaflet.