14.4 How to write good code/workflow!

• Q: What is your experience with looking at data analysis code you have written 2 years earlier?

1. Comment your code
   • Data sources, data manipulations, describe steps of analysis
   • commenting increases readability
   • commenting increases reproducability
   • you can’t comment too much
   • Comments in R: #
   • Use meaningful names!
   • Don’t be too clever!
2. Structure you code properly
   • Headlines
   • Indentation (https://google-styleguide.googlecode.com/svn/trunk/Rguide.xml)
   • Write robust syntax (e.g. relative paths)
   • Follow the Google Style Guide for R https://google-styleguide.googlecode.com/svn/trunk/Rguide.xml
3. More rules
   • Code must be usable on machines with different paths to the project
     • Path is set once and afterwards, relative paths are used
     • There is a folder structure separating data and figures (and code)

For larger projects you might consider tools that keep track of versions, branches etc. like Git (http://www.github.com among others offers free git repositories). github can easily be used from within RStudio.

3. More…
   • See tips by Jonathan Nagler
   • Try the formatR package: Online example
   • tidy_dir(): Formats all R files in directory