Abstract for EDEG 2018 submission

These files contain the analysis and abstract for the European Diabetes Epidemiology Group 2018 conference.

Results are found in data/pls_*.rda. If the original data is available, can regenerate the results using:

devtools::load_all()
generate_pls()

Poster can be generated by running the command:

generate_poster()

Poster can be found at: https://doi.org/10.6084/m9.figshare.6159293

Introduction to this project

This README details how this research directory is structured, how files should be run, and what the different files do. The layout and setup of this project was designed for using RStudio and devtools. It is set up this way to make it easy for others to run your code and analyses for themselves and to skaffold onto devtools (used for R package development) because it is well documented and actively maintained. See the excellent R for Data Science online book for more details on how to work with this directory format.

Typical commands used in this workflow include:

• Ctrl-Shift-L (devtools::load_all())
• Ctrl-Shift-D (devtools::document())
• Ctrl-Shift-K (rmarkdown::render('file.Rmd'))
• devtools::use_package('packagename') (when using another R package)

Using and installing other packages

There is no need to use library() functions when using a package. Use devtools::use_package('packagename'), followed by packagename::function_name() when using the functions. This approach allows you to make use of devtools features of using and installing the necessary packages.

To install the dependency packages for this project, simply use:

devtools::install(dependencies = TRUE)

General folder details

The project directory is generally structured with the following folders:

• Base folder (./)
• R/
• vignettes/
• doc/
• data/ (optional)
• .git (optional)

Base (parent) folder

Contains a few files:

• .gitignore tells Git to ignore certain files from being tracked and prevents them from entering the version control history.
• .Rbuildignore tells devtools which files to not include when running functions such as devtools::load_all() or devtools::install().
• DESCRIPTION is a standard file included that allows devtools to run it's functions, which in turn make your life easier for running analyses and writing up your results. It provides a description of what the project does and most importantly what R packages your project relies on.
• NAMESPACE is also standard for devtools and is used more for others when they view your code and analyses.
• .Rproj is the file to dictate that the directory is a RStudio project.

R/ folder:

Contains functions and code used by all subsequent .R or .Rmd files and can be accessed by documents in the vignettes/ by using devtools::load_all(). There are at least four files (probably more):

• fetch_data.R to get, process, and save the dataset.
• load_data.R loads or updates (if fetch_data.R has been changed) the dataset in the data/ folder.
• setup.R to run options for the packages.
• functions.R to hold all custom functions used for the analysis.
• zzz.R holds functions or code that will run automatically after using devtools::load_all().

doc/ folder:

Contains the main product of the project: the slides, manuscripts, or other final products.

vignettes/ folder:

Contains (at minimum) the main document file that will present the results of the analysis and likely also other files that may supplement the main document.

data/ folder (optional):

The data folder contains the analysis-specific dataset. Meaning this dataset may be a subset of an original dataset, keeping the data relevant to the research question without keeping the (potentially) larger dataset around.

Resource

Since this layout is based on R package development, check out the online book on R package development to learn more about how to make the most use out of this project layout (and why prodigenr structures it this way).