Using the shiny app

All content below relates to the app/ folder.

What this does

Infer the number of circulating COVID-19 cases from newly reported deaths. This is expected to be useful in places where surveillance has not picked up cases yet, and new deaths are the only indication that there are more circulating cases.

How to use it

Packages needed

The following instructions will install the packages needed if they are not present on your system:

if (!require(pacman)) {
  install.packages("pacman")
  library(pacman)
}
p_load("shiny")
p_load("ggplot2")
p_load("magrittr")
p_load("incidence")
p_load("projections")
p_load("distcrete")
p_load("DT")
p_load_gh("reconhub/reportfactory")

Note that you only need to do this once (or not at all, if all these packages are already on your system).

Starting the app

To start the app, open R inside this folder and type shiny::runApp("death_to_case.R"). If using Rstudio, make sure you view the app within a decent web browser by clicking on "Open in Browser".

Using the 'analyses' reportfactory

This folder contains a reportfactory with reports providing a proof of concept and implementation of the model used for predicting cases from recent COVID-19 deaths.

Initial setup

You will first need to install dependencies before compiling the documents in this factory. We recommend using the latest version of R. Double-click on open.Rproj (or just start R in the model/ folder) and copy-paste the following instructions:

if (!require(reportfactory)) remotes::install_github("reconhub/reportfactory")

library(reportfactory)
rfh_load_scripts()
install_deps()

Compiling the reports

To compile the report, double-click on open.Rproj (or just start R in the analyses/ folder) and type:

reportfactory::update_reports(clean_report_sources = TRUE)

Dated and time-stamped outputs (including the html version of the report) will be generated in report_outputs.

Using the source

The source of the report is an Rmarkdown document stored in report_sources/. If you plan on working on your own local copy, we recommend either using version control systems (e.g. GIT) to track changes, or creating new versions in report_sources/ with a more recent date in the file name. Note that by default, reportfactory::update_reports() will always compile the latest version of reports.

Disclaimer

This is work in progress, which has not been peer-reviewed yet. Do not use without consulting me before.

Distribution

• Code work is distributed under MIT License (copyright: Thibaut Jombart 2020).
• Other documents (analyses/) are distributed under CC-BY