index.R

#' Install and load dependencies
#' - make sure you have pacman installed: install.packages("pacman")
#' - or install and load manually
pacman::p_load(data.table, survey, ggplot2, patchwork, magrittr, kableExtra, stringr, lmtest, zscorer, mitools, viridis)

#' Set the analysis directory (usually the root directory of repo)
analysis_dir = getwd()

#' Should tables be created (set to FALSE if correct software is not installed)
make_table = TRUE

#' What variables should be used to generate the post-stratification weights?
#' - see script 2_prepare_data.R for more details
#' - Note. Strata become too sparse when stratifying by age, sex, and householdsize
#'   POSTSTRATIFICATION_STRATA = "age_sex_householdsize" will result in errors
POSTSTRATIFICATION_STRATA = c("age_sex_householdsize", "age_householdsize", "age_sex", "age", "none")[3]

#' Set to TRUE to skip other analyses
CONTACT_DATA_ONLY = TRUE
#for(POSTSTRATIFICATION_STRATA in c("age_householdsize", "age_sex", "age", "none")){
  #' Set global options, create output folders, load helper functions, and load data
  source(sprintf("%s/scripts/0_setup.R", analysis_dir))
  
  #' This estimates the total population size in Digaale, and FPC to use in the models
  #' - creates table_sB1_vacant_households
  #' - creates table_sB2_households_visited
  source(sprintf("%s/scripts/1_fpc_correction.R", analysis_dir))
  
  #' This does some additional cleaning of the data, assigns age-groups, creates new variables, and creates the survey
  #' objects to use when correcting for FPC and to do the poststratification
  source(sprintf("%s/scripts/2_prepare_data.R", analysis_dir))
  
  #' Create table 1: Characteristics of participating households and prevalence of risk factors in Digaale IDP camp
  #' - creates table1_characteristics_and_risk_factors
  if(!CONTACT_DATA_ONLY) source(sprintf("%s/scripts/3_table1.R", analysis_dir))
  
  #' Create tables 2, 3, and supplemental table C1
  #' - creates table2_travel
  #' - creates table3_contacts
  #' - creates table_sC1_school_work_sens
  source(sprintf("%s/scripts/4_tables_other.R", analysis_dir))
  
  #' Regression analysis assessing collected variables with self-reported pneumonia incidence
  #' - creates table_sD1_regression_pneumonia
  #' - creates table_sD2_regression_pneumonia_6m
  if(!CONTACT_DATA_ONLY) source(sprintf("%s/scripts/5_tables_regression.R", analysis_dir))
  
  #' Create figure 1: Demographic distributions in Digaale IDP camp
  #' - creates figure1_demographics
  if(!CONTACT_DATA_ONLY) source(sprintf("%s/scripts/6_figure1.R", analysis_dir))
  
  #' Create figure 2: Contact frequencies, types, and matrices
  #' - creates figure2_contacts
  source(sprintf("%s/scripts/7_figure2.R", analysis_dir))
  
  #' Additional analyses contact estimates
  #' - creates figure_sC1_bootstrapped_matrices
  #' - creates figure_sC2_intra_extra_household_contacts
  #' - creates figure_sC3_contact_bysex
  #' - creates figure_sC4_household_contacts_expected_reported
  source(sprintf("%s/scripts/8_figures_supp.R", analysis_dir))
#}

CONTACT_DATA_ONLY = TRUE
contacts_sensitivity_data = list()
for(POSTSTRATIFICATION_STRATA in c("age_householdsize", "age_sex", "age", "none")){
  #' Set global options, create output folders, load helper functions, and load data
  source(sprintf("%s/scripts/0_setup.R", analysis_dir))
  source(sprintf("%s/scripts/1_fpc_correction.R", analysis_dir))
  source(sprintf("%s/scripts/2_prepare_data.R", analysis_dir))
  source(sprintf("%s/scripts/9a_contacts_stratum_data.R", analysis_dir))
}
source(sprintf("%s/scripts/9b_contacts_stratum_figure.R", analysis_dir))