Reproducible research is part of several large trends

• Openness
• Transparency
• Quality: Reproducibility
• Collaboration / Team work
• Communication

Many rapid changes to key infrastructures of science

We are in the middle of an exponential growth curve:

• Data production
• Data storage and transfer
• Computing power
• Published research
• Complexity of methods

Good modern example: UK Biobank

But we scientists still think from "cottage industry" perspective

• Industrialisation of the research work flow
• Specialisation in research tasks

Why? Still very strong barriers to change

• Tools needed
• Tradition, culture and common practices
• Researchers don't yet see value in adopting open, reproducible workflow
• Training and reward systems need to be adapted:
  • Publication
  • Academic recognition / careers
  • Research funding mechanisms
• Law: privacy concerns about sharing data, IP protection, patents, etc.

Little to no training in software or programming

Source from xkcd.

What does it mean for you?

• Need to stay updated: Keep learning
  • Read bookdown books (most free)
  • Ask (and answer!) questions on StackOverflow
  • Teach others!! Great way to learn :)
  • Read e.g. Open Science Handbook

• Find and collaborate with those familiar with these concepts (online and/or in real life)

• Cite research that is or tries to be more reproducible

• Keep the principles of reproducibility in mind, then find the tools

• Practice reproducible and open science

  • More on this later in session

Recognize importance of code and data: Cite them!

# Example:
citation("dplyr")

## 
## To cite package 'dplyr' in publications use:
## 
##   Hadley Wickham, Romain François, Lionel Henry and Kirill Müller
##   (2019). dplyr: A Grammar of Data Manipulation. R package version
##   0.8.1. https://CRAN.R-project.org/package=dplyr
## 
## A BibTeX entry for LaTeX users is
## 
##   @Manual{,
##     title = {dplyr: A Grammar of Data Manipulation},
##     author = {Hadley Wickham and Romain François and Lionel Henry and Kirill Müller},
##     year = {2019},
##     note = {R package version 0.8.1},
##     url = {https://CRAN.R-project.org/package=dplyr},
##   }

Comment: True reproducibility is very difficult

• Requires self-contained virtual environment
• With exact package versions and operating system used
• Tools to do this include:
  • Docker virtual containers [1]
  • Continuous intergration with [Travis] [2]

...but this is not the goal, nor should it be

• Tools to simplify this are being developed
  • Keep eye out