README.md
In healthiar: Quantifying and Monetizing Health Impacts Attributable to Exposure

Test data

Test-only inputs and expected results for the testthat suite.

Always with the namespaced reader plus testthat::test_path(), two arguments, so that every call site looks the same:

data   <- readRDS(testthat::test_path("testdata", "airqplus_pm_copd.rds"))
expect <- utils::read.csv(testthat::test_path("testdata", "expected_exposure_pwm.csv"))
pop    <- terra::rast(testthat::test_path("testdata", "population.tif"))

Names are lowercase snake_case, prefixed by topic (noise_, ozone_, erf_, pop_, lifetable_, expected_).
Inputs are .rds (xz-compressed); expected results stay .csv so that changes to reference values are visible in a diff.
Plain data.frame / tibble only, so that the test suite needs no package beyond the ones in DESCRIPTION.
No exdat_ prefix here: that prefix is reserved for the package datasets in data/.

Data that is part of the package interface is used directly, never copied into this directory:

| Data | Where it lives | How to use it in a test | | ------------------------- | --------------- | ------------------------------------------------------------------------------ | | exdat_* (tibbles, sf) | data/ | exdat_pwm_2 (lazy-loaded, no prefix needed) | | exdat_pwm_1 (GeoTIFF) | inst/extdata/ | terra::rast(system.file("extdata", "exdat_pwm_1.tif", package = "healthiar")) |

The raw sources and the scripts that build both are in data-raw/, which is excluded from the built package via .Rbuildignore.

Two data sets hold an AirQ+ life table assessment of the mortality attributable to PM2.5 in Switzerland in 2019, each with the input data as entered in AirQ+ (input, pop) and the results of AirQ+ (output), so that both tools can be compared on exactly the same data:

| Data set | Exposure | AirQ+ export in data-raw/ | | --------------------------------- | ---------------- | -------------------------------------------------------- | | airqplus_pm_deaths_yll.rds | constant | exdat_lifetable_airqplus_deaths_yll_lifetable_adults.csv | | airqplus_pm_yll_single_year.rds | one single year | airqplus_pm_yll_single_year.csv |

Both were parsed from those exports, which are kept as the reference. airqplus_pm_yll_single_year.rds is built by data-raw/prepare_airqplus_export.R, whose function reads either export. The 0 deaths of the females aged 8 were replaced by 1, because a survival probability of 100% has no conceptual logic and healthiar warns about it; this does not change the results of AirQ+ at the two decimals that it exports.

Compare with the male and the female results of AirQ+ and never with its "All genders" ones: the latter are not the sum of the former but a life table with both sexes pooled. See the chapter "YLL & deaths with life table" of the vignette.