This code reproduces the combined public-private sector figures from the manuscript https://www.medrxiv.org/content/10.1101/2022.07.22.22277932v1.full.

The easiest way to use this code is via the GNU make build tool; the code can also be run directly from R.

Running this code:

Download data

• Time series data for running this code can be downloaded from: https://zenodo.org/record/6948468.

• Filenames include a province code. Each data file includes 25 imputated timeseries (see manuscript for description of imputation procedure).

• Data files should be put into a folder named 'data'.

• the R code provided runs out the box using the .RDS data files; .csv files are also provided

Using make:

• run 'make setup' to set up directory structure
• run 'make' to generate R estimates and plots with 7-day sliding windows
• run 'make plots_<window_size>' (e.g. 'plots_14') for generate R estimates and plots with your desired length of sliding window.
• run 'make all_plots' to generate R estimates and plots with 7, 14, and 21-day sliding windows.
• set variable "UNCERTAINSI" to TRUE if you want to reproduce figures in the manuscript (using a less-certain serial interval), or set to FALSE for faster runtime.
• set variable "PERIODANALYSES" to TRUE to generate R estimates per lockdown level, or FALSE to skip this.
• set variable "NCORES" to specify how many of your machine's cores you'd like to allocate to this task.
• other variables (e.g. plot resolution or level of sampling for uncertain serial interval

Using R without make:

• Set up repository structure:

  • data/
  • results/tables
  • results/Period_analyses
  • plots/

• Variables should be edited inside the "if(interactive()){c(<variable_1>, <variable_2>, etc.)}" portion of the ".args <-" assignment block. See comments and/or code just below the assignment block for which variable is which.

• To produce R estimates:

  • run "est_R_hosp.R" with the second argument set to "admit" (for estimates based on hospital admissions)
  • run "est_R_hosp.R" with the second argument set to "deaths" (for estimates based on hospital-associated deaths)
  • run "est_R_lab.R" (for estimates based on rT-PCR-confirmed cases)

• To produce plots:

  • first generate R estimates with the desired window size
  • run "plot_results.R" with the first variable set to the desired window size