CORD-19 Data Package
R package development and documentation-writing
Notable topics: R package development and documentation-writing
Recorded on: 2020-03-18
Timestamps by: Alex Cookson
Screencast
Timestamps
Overview of JSON files with the data David will make a package of
Starting to create a new package with "New Project" in RStudio
Creating a file to reference the license for the dataset
Using use_data_raw function from usethis package to set up a folder structure and preliminary function for raw data
Explanation that we want to limit the number of packages we load when building a package (e.g., no library(tidyverse) )
Using use_package function from usethis package to add "Suggested packages"
Reviewing import and cleaning code already completed
Using roxygen2 package to write documentation
More documentation writing
Using use_data function from usethis package to create a folder structure and datafile for (finished/cleaned) data
Making a mistake clicking "Install and Restart" button on the "Build" tab (because of huge objects in the environment) (see 26:50 for alternative)
Using load_all function from devtrools package as an alternative to "Install and Restart" from above step
Using document function from devtools package to process written documentation
De-duplicating paper data in a way the keeps records that have fewer missing values than other records for the same paper
Using use_data function with its overwrite argument to overwrite existing data
Writing documentation for paragraphs data
Testing an install of the package
Adding link to code in documentation
Writing examples of how to use the package (in documentation)
Discussion of outstanding items that David hasn't done yet (e.g., readme, vignettes, tests)
Creating a simple readme, including examples, with use_readme_rmd function from usethis package
Using knit function from the knitr package to knit the readme into a markdown file
Creating a GitHub repository to host the package (includes how to commit to a GitHub repo using RStudio's GUI)
Explanation that version 0.0.0.9000 means that the package is in early development
Actually creating the GitHub repository
Overview of remaining tasks