All COVID19 graphs and maps. Other R stuff that's public moved to r_play.

Note that rgeos and maptools R packages were depreciated for newer versions of R.
Tests in github workflows now fail because rgeos can't be fetched.

I created these graphs and maps both as a learning excercise and a way to get custom graphs to visualize what I wanted too. I also didn't care for the way many US maps didn't include Alaska, Hawaii, and the major US territories. So, I made insert maps for those as well as an insert for DC which is often invisible.

The site for my graphs and maps is https://sites.google.com/site/salishseacam/covid19

To run from the command line without adding environment variables: Rscript covid19_run.R

It can be modified it to make the graphs others are interested in.

CSV files of the data used are published along side the graphs

The raw data used for covid19 cases is from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. To simplify I used the population data provided in the same project. Vaccination data is also pulled but not used much since it's so spotty.

For maps census bureau shapefiles are used since they contain more of the US but maps from map_data are also used for consistent borders. Data continuity issues are introduced when data from inconsistent sources is joined. Business logic for that is regretabbly sprinkled throughout and should be re-organized.

The R code is broken into a file of functions and a much smaller script to make the graphs. The covid19 functions should certainly be made into a package. I'd played with that but it got complicated with some of the testing.

There is certainly some dubious content in here. That is part of the learning process. Comments and suggestions are always welcome. Or fork and send pull requests if you're git literate.

There is some dead-ended threads in the code. Sometimes slicing and dicing the data in dubious way. One such mostly abandoned approach was visualizing cases by bifurcating based on regions voting patterns.

My products are pushed into AWS S3 buckets. The keys and bucket name are set in a command line wrapper script, an example is included. If you wanted to use S3 buckets you could set the environment var somewhere else or do a similar wrapper for however your running R.

Administrative Levels

There is an effort to be more international and less US centric. So, insteat of States / Counties we're trying to use Admin Level 1 and Admin Level 2. For now there is a mix.

Levels break down like this: country admin1 e.g. us state, cnd province admin2 e.g. us county, french departments

Stuff in here: LICENSE - Apahce license README.md - this file Rwrapper.sample - sample wrapper covid19_functions.R - the bulk for R code covid19_install_packages.R - script for installing all libraries used covid19_map_run.R - just make maps covid19_run.R - make the whole kit and kaboodle html_table.py - make table of states and counties run_tests.sh - for testthat tests/ - directory for github actions tests work_in_progress/ - mostly cruft

Notes on tigris To enable caching of data, set options(tigris_use_cache = TRUE) in your R script or .Rprofile.

Work to do:

• refactor making maps
• redo maps using something snazzy like leaflet for interactive maps
• clean-up warnings from map making (probably suppressing wanrings on duplicating projections)
• add more tests
• clean-up or remove vaccination data
• built_all_states - make wa_and to be more generic
• pull all of the configuration of making graphs and maps out and put in separate file. maybe even make it explicit configuration.
• make it a package
• maybe used the new york times curated data?