serosurvr

Using serosurvr

Installation

install.packages("devtools")
devtools::install_github("serotracker/serosurvr")

Environment Variables

If you are a SeroTracker developer, researcher, or partner, or CITF modelling team member that needs to pull new data, add a .env file to the top level of the repository to store environment variables. This file should be formatted as follows:

RECORDS_URL=
FILTER_OPTIONS_URL=

Check the pinned message in the #data-modelling Slack channel or ask Harriet for the environment variables that you will need.

Setting file options

Set import_new_data = TRUE to import new data or import_new_data = FALSE to use cached data.

Set the path to cache folder to read from and write to for this project, e.g. path_to_cache_folder = ".cache/example_analysis"

Creating a data request

A basic data request has parameters reqname and estimates_subgroup. The estimates_subgroup parameter takes 3 values:

• all_estimates: return all estimates [DEFAULT]
• primary_estimates: return only primary estimates (one summary estimate per study identified by SeroTracker)
• prioritize_estimates: return only prioritized estimates according to prioritize_estimates_mode

If using prioritize_estimates, prioritize_estimates_mode takes 3 values related to test adjustment:

• dashboard: Prioritizes study author's test adjusted ("dashboard primary") estimate.
• analysis_static: Prioritizes study author's test unadjusted ("academic primary") estimate.
• analysis_dynamic: Prioritizes SeroTracker's own test adjusted estimate. If we were unable to successfully adjust the seroprevalence estimate ourselves, returns the study author's test adjusted estimate. [DEFAULT]

The full estimate prioritization code is stored here: https://github.com/serotracker/iit-backend/blob/b5dfe5c8af42f652fded59303ab97d04847be16c/app/utils/estimate_prioritization/estimate_prioritization.py

Example:

fs_males_req <- serosurvr::datareq_params(reqname = "data_req",
                                          estimates_subgroup = 'prioritize_estimates',
                                          prioritize_estimates_mode = "dashboard")

Requesting data

To request data, pipe the data request to the get_data function.

Example:

tbl <-
  data_req %>%
  serosurvr::get_data(import_new_data = import_new_data,
                      path_to_cache_folder = path_to_cache_folder)

Improving serosurvr

Pushing new versions of serosurvr

Before pushing a new version of serosurvr, please follow the following steps:

• Run devtools::document() to ensure that the package is fully documented
• Run devtools::check() to ensure that all checks pass and the updates are ready to push

renv

If you are modifying this library, please note that it relies on renv for package management.

To set up renv for this repo, follow the following steps:

• Clone this git repo
• Ensure you have the renv package installed (in R, run install.packages("renv"))
• Start R in the sysreview_biostats working directory, either by opening sysreview_biostats as a RStudio project or launching R from your console.
• Call renv::restore() to restore the project library

If you add a new package or dependency to this repo, make sure you log this change by calling renv::snapshot() and then committing changes to the renv.lock file. See this page for more information on the renv workflow.

The collaborating with renv page provides further information on how to use renv with Git.