README.md
In the-pudding/puddingR: Pudding Helper Functions

puddingR

puddingR is designed to streamline R data analysis and open data release for the team at The Pudding. Check out the full pkgdown-generated site here.

Learn more by searching vignette("puddingR") in your R or RStudio console.

At the moment, there are no plans for puddingR to appear on CRAN. To download:

# Install development version from GitHub
devtools::install_github("the-pudding/puddingR")

# Load into your R environment
library(puddingR)

The puddingR package currently has a few main features:

For our front-end work, the team at The Pudding, uses a starter template which sets up anything that could be useful for our work and creates a file directory system. The directory helps the team (and the author) know exactly where to find everything they need for a project.

The create_project() function does the same thing for R projects. It creates a single folder, by default called analysis, creates a new R project and several directories. The resulting structure looks like this:

analysis
|-- assets
  |-- data
    |-- open_data
    |-- processed_data
    |-- raw_data
|-- packrat
|-- plots
|-- R
|-- reports
|-- rmds
|-- analysis.Rproj

Many of the remaining functions in this package use default values that assume the above folder structure is present.

puddingR::create_project(title = "My Cool Project")

While RStudio comes with several themes to use for Rmarkdown reports, I have created one specifically for use with Pudding projects. It contains our logo and lots of tips for how to import and display data.

To use it, in RStudio, navigate to File > New File > Rmarkdown

A window will pop up with a list of options on the left-hand side. Select From Template from that list, and navigate to Pudding Report. This will launch open a new Rmd file filled with tips that can be deleted if you don’t need them. If you’re using the above file structure, save this into the rmds folder.

Often in our work we need to write data that we have analyzed and save it in another location. In this workflow, most processed data belongs in assets/data/processed_data/. The easiest way to get it there is to use export_data(). This function has several arguments, included in them is location. If you are using the default file structure from puddingR, you can simply use open to save it to the open_data directory, processed to save it to the processed_data directory, or, if this analysis folder exists in the same working directory as the front-end Pudding starter template, you can use js (which will save the file in src/assets/data/ for use in front-end work).

# To export mtcars dataset to the processed_data directory
puddingR::export_data(mtcars, "cars_data", location = "processed", directory = "auto", format = "csv")

If you want to export codebooks and the scripts used to generate data at the same time the data is generated, you can use export_all() instead.

# To export mtcars dataset to the open_data directory
puddingR::export_all(mtcars, "cars_data", 
                      location = "open", 
                      directory = "auto", 
                      format = "csv", 
                      # Also export a codebook to accompany the data  
                      codebook = TRUE,
                      # Export the code chunks from an Rmd file used to generate the data
                      scripts = c("load_data", "analyze_data", "graph_data"),
                      scriptFile = here::here("rmds/analysis.Rmd")
                      )

These pieces can be accomplished separately using export_data(), create_codebook() and export_code().

The default directory for a codebook is in open_data/intermediates. These files should be opened and manually edited to provide the correct metadata to accompany an open data file.

Once all of the intermediate codebooks have been edited, they can all be combined into a single file using knit_data_readme().