This vignette demonstrates the washr data package
development workflow with a toy dataset titled fssample.
The fssample data is an (imaginary) 5-day sample collection
of faecal sludge.
You are taking 20 samples of 1 liter faecal sludge from pit latrines and septic tanks at households and public toilets (5 samples each). For each sample, you note the number of daily users of the sanitation system.1
1. Create and process dataset
After initializing an R package named with fssample with
devtools, you can start to set up the raw data in the
package by executing:
Under the fssample directory, there is now a new
directory data-raw with an R script
data-processing.R inside it. Go to
data-processing.R and refer to our template code to import,
clean, and export the dataset. For instance, you may want to change the
data type of the column “location” to be factor or reformat the column
“date” into YYYY-MM-DD.
After executing the last a few lines in
data-processing.R, namely,
usethis::use_data(fssample, overwrite = TRUE)
fs::dir_create(here::here("inst", "extdata"))
readr::write_csv(fssample,
here::here("inst", "extdata", paste0("fssample", ".csv")))
openxlsx::write.xlsx(fssample,
here::here("inst", "extdata", paste0("fssample", ".xlsx")))A directory data/ that contains the exported data in
.rds format was created in the root directory.
2. Document dataset
The next step is to provide human and machine-readable documentation for the dataset and the package itself.
For documenting the package, you work with the
DESCRIPTION file by running:
Next comes the dataset documentation. You first create a dictionary for the dataset in CSV file format.
Go to data-raw/dictionary.csv, open the CSV file, and
fill in the empty column description of the dictionary.
Once the dictionary is complete, you document the dataset by turning the
content of the CSV file into roxygen comments by executing:
Go to R/ and fill in the title and description for the
dataset.
Communicate dataset
The R dataset and documentation are complete. It’s time to communicate with the public using human-readable and visually appealing tools. We currently achieve this with the following two components.
- README
- pkgdown website
setup_readme()
# Go to README.Rmd and complete this R Markdown file
build_readme() # Generate README.md from README.Rmd
setup_website()Now, it’s time to work on polishing the README and website. Once you
are satisfied with them, don’t forget to re-run
build_readme() and build_site() again to
update.