ProjectTemplate is an R package that makes it easy to create and run projects so that you can spend less time thinking about organization and more time analyzing data.
3. Things I care about (in structuring projects)
+ Easy to find things (reports, data, etc.)
+ Easy to share
+ Ability to update with new data
+ Reproducibility
+ Being able to quickly verify results
+ The sanity of my future self
5. Technical Benefits of ProjectTemplate
+ Easy to create new projects
+ Organizes and standardizes your projects
+ Automatically loads data and R packages needed
+ Automatically runs data munging scripts
6. Non-technical benefits
+ Automates the thoughtless parts of your project (so you
can use your energy on the important stuff)
+ Easier for someone new to your project to understand
steps / data / preparation involved to execute it
10. Advice
+ Number files in the munge folder so they run in order
+ Avoid manually updating raw data files (do this in code!)
+ Edit the global.dcf file (in config folder) to add libraries you
use, load libraries automatically, and adjust “strings as
factors” settings
+ Lots of other ways to make analysis reproducible;
ProjectTemplate is one part of larger ecosystem
11. Cheat Sheet
ProjectTemplate home page
+ Don’t miss the tutorial!
Great discussion on best practices for managing analysis projects
The talk that got me interested in ProjectTemplate (thanks, Hilary Parker!)