Introduction to reproducibility

Julia Romanowska

2023-10-24

Outline

• Overview of the book
• Structure of the meetings
• What is reproducibility?
• Why do we need reproducibility?
• Project start (based on chapter 3)

Go to mentimeter first!

Photo by Marion Botella on Unsplash

Overview of the book

Part 1:

• version control
• literate programming
• functional programming

Part 2:

• freezing versions
• packaging code
• testing
• automation
• encapsulation

Structure of the meetings

Proposed structure:

• short presentation
• exercises/hands-on/live coding
• discussion

Presentation summarises the part from the book.

Use exercises from the book and/or some of your own!

Lastly, let’s have time to talk, ask questions, ask for help on your problem…

We need your participation!

You can form teams, so that you will need to only prepare a very small part!

Before we start…

What is reproducibility?

Sharing:

• analysis steps
• inputs, outputs
• documentation: data + code
• code used to generate each figure
• notes
• versions of software, packages, etc.

Ask first and wait for answers.

Why do we need reproducibility?

Ask first and wait for answers.

• do you know how to get to each single figure in the paper you published?
• do you know what is the content of the file X in a folder you created a year ago?
• would it be easy to continue on your/your colleague work?

Fundamentals

Prerequisites

• R is not RStudio
• text file vs. binary file
• R packages - installation and use
• variables/data in the environment (memory)
• file location, relative and absolute paths
• file extension

variables/data in the environment:

• check it in “Environment” tab or with ls() command

• don’t save the image when quitting R:

  • use R --vanilla or
  • change in Global settings in RStudio

file extension:

• the part of the file name after the last .
• can be changed by the user - no magic needed!
• be careful: this should indicate what type of data is in the file
• R-related data: .rda, .rds, .RData

Project start

Analysis of data: Housing in Luxembourg

• data: prices of housing in Luxembourg,
  nominal prices,
  per commune,
  trapped in Excel

• write scripts:
  rescue data from Excel,
  convert to real prices,
  make some plots/tables

Hands-on

Your project is not done

• How easy would it be for someone else to rerun the analysis?
• How easy would it be to update the analysis once new data gets published?
• How easy would it be to reuse this code for other projects?
• What guarantee do we have that if the scripts get run in 5 years, with the same input data, we get the same output?

• project documentation + pipeline management + {renv}
• testing!
• structuring into functions and packaging
• freezing the code: {renv}, Docker

Photo by Joshua J. Cotten on Unsplash

Photo by Dewang Gupta on Unsplash