This is the CASES data visualization project. Original manuscript here.

Overview

The CASES app has two major sections - data generation and data visualization. The data generation is done in Julia, but the data is parsed in, and preserved by, the Python code.

The initial run will execute the Julia code on each sector, as specified in derived_data_inputs.tsv and employment_input_data.tsv.

Initial run will take quite some time (> 1hr). Once these data are generated, launching will take only a few seconds.

Production runs should be done in the Docker environment.

Inputs

WARNING: Commented out headers matter - the month info is important in employment_by_sector_by_month Apologies - these input formats are FRAGILE. Case sensitive, column name sensitive, the works. Handle with care.

Input file	Description
derived_data_inputs.tsv	Master file. The first line lists the directories containing the employment file path, the age fracs path, and the employment percentage in feb 2020
../employment_by_sector_Feb2020	Starting point data. Path referencing derived_data_inputs.tsv.
../age_fracs	Per SES age per sector age breakdown percentages.
r_input_data.tsv	Specifies the R actuals by SES index number.
../R	Per SES R actuals refrenced by r_input_data.tsv.
employment_input_data.tsv	First line points to the containing directory, subsequent lines contain SES index number and actual employment by SES.
../employment_by_sector_by_month	Employment actuals referenced by employment_input_data.tsv
../mortality	Mortality by SES time series. Not used, included for completeness.
sector_color_mappings.tsv	Color values assigned to each sector. Taste the rainbow!
sector_names.tsv	Maps sector name as expressed in data to the display sector name

Running in docker - scripts

build.sh will create the environment.

Delete.sh does what it says on the tin - removes the Docker environment

run.sh runs the production environment.

exec.sh will attach to a running docker so you can run in a shell.

docker_start.sh is used internally during dockerized running.

Initial run

Build the system with "build.sh". These are standard docker commands, and can be run manually on windows systems.

Run with "run.sh". This will start the system but won't kick off intermediate file generation until the web url is hit (url wil be http://127.0.0.1:80 ). However, for better progress reports, it's easier to run "exec.sh".

# ./exec.sh
root@de13780a0da3:/var/www/apache-flask# cd app
root@de13780a0da3:/var/www/apache-flask/app# python3 ./CASES.py

This will take some time to run and will generate the intermediate .bin files.

At this point, it's safe to exit the prompt. The systems should be running on localhost port 80. If not, "delete.sh" "build.sh" and, finally, "run.sh".

All done!

---

Manual setup - NOT REQUIRED

This is for development only; the entire system can and should be run in docker.

Set up Pyjulia

Getting pyJulia setup requires a bit of work, and it will require upkeep if your python3 environment changes signifigantly.

First, install python and setup your pyenv:

python3 -m venv env
activate ./env/bin/activate

Next, install julia from (here)[https://julialang.org/downloads/]

Next, install pyJulia from (here)[https://pyjulia.readthedocs.io/en/latest/installation.html] (don't forget to run the "julia.install()" step)

Add julia to your path. On the mac, that's going to look like this (add this line to your .bashrc):

export PATH="/Applications/Julia-1.5.app/Contents/Resources/julia/bin:$PATH"

Finally, run julia and type:

import Pkg; Pkg.add("DataFrames")
import Pkg; Pkg.add("DifferentialEquations")
import Pkg; Pkg.add("CSV")

control-d to exit.

Set up and run the python environment

Create a python environment:

python3 -m venv env
source ./env/bin/activate
python3 -m pip install -r requirements.txt
./cases.py

Navigate to: http://127.0.0.1:8050/ per the output

Notes and bugs

Docker

Docker will require more than the default 2gb that's assigned on Macs (Linux docker runs out of the box.)

Note that docker runs as user www-data, so set permissions accordingly on visualization dir for deployment

Notes 6/1/2021 added US and ca:

create mode 100644 employment_by_sector_by_month/CA_11.tsv create mode 100644 employment_by_sector_by_month/US_12.tsv

create mode 100644 age_fracs/US_age_fracs.csv create mode 100644 employment_by_sector_Feb2020/CA_employment_by_sector_Feb2020.csv create mode 100644 employment_by_sector_Feb2020/US_employment_by_sector_Feb2020.csv

create mode 100644 r/CAstate/R_CA.csv create mode 100644 r/R_US.csv