I will first

  • introduces ways to organise files on your computer and to document your workflows.
  • how to make your code more reproducible and readable
  • how to write reproducible reports.
  • the basics of version control.
  • how to archive your code.

Organising projects for reproducibility

I embrace a one folder = one project mentality. Rstudio's "R projects" are excellent for encouraging this. This habit enables easy communication with other project stakeholders and for that reason it is so important.

Make sure to use ISO 8601 format (YYYY-MM-DD) to avoid confusion between differing local dating conventions.

Choose filenames so that alphabetical sorting will organise types of files for you: in the same way that ISO 8601 puts the most significant unit (year) first so that alphabetical = chronological.

Avoid using absolute paths and use only relative paths. An absolute path is one that gives the full address to a folder or file. A relative path gives the location of the file from the current working directory.

For example:

# Absolute path -----------------------------

# Relative path ------

When working in R, RStudio projects can help aid portability of code. RStudio projects provide a self-contained coding environment to work in, and when opened, a `.Rproj` file sets the project working directory and saves the history and state of the project. RStudio projects can be created directly from the RStudio GUI and can be created in an existing folder, a new empty folder, or checked out from a version control repository.

Here is a basic project directory structure:

  • The data folder contains all input data (and metadata) used in the analysis.
  • The doc folder contains the manuscript.
  • The figs directory contains figures generated by the analysis.
  • The output folder contains any type of intermediate or output files (e.g.simulation outputs, models, processed datasets, etc.). I also might separate this and also have a cleaned-data folder.
  • The R directory contains R scripts with function definitions.
  • The reports folder contains RMarkdown files that document the analysis or report on results


Writing functions

how to define your own functions in R and in Python:

# Create a user-defined function in R:
square_number <- function(base){         
  square <- base*base         

# Calling a user-defined function in R
# Create a user-defined function in Python:
def square_number(base):      
     square = base*base      

# Calling a user-defined function in Python:
x_square = square_number(5)

Show the packages used

With R, the simplest (but a useful and important) approach to document your dependencies is to report the output of sessionInfo() (or devtools::session_info()). Among other information, this will show all the packages and their versions that are loaded in the session you used to run your analysis. If someone wants to recreate your analysis, they will know which packages they will need to install.

Conducting a reproducible analysis

My reproducible project workflow enables me to answer

  • What did I do?
  • Why did I do it?
  • How did I set it up?
  • When did I make changes?
  • Who needs to access it?