1. Don't reinvent the wheel/package
One of the main reasons to use R is the number of freely
available packages.
2. Packages everywhere!
From around 80 or so in 2005, we are now approaching 15000! At times, it seems there is a package for everything.
But how do you choose which package to use?
3. Choosing an R package
For the most part, common sense takes you quite a long way.
For example, is the package mature? How old is the package?
Is the package being actively developed, or was the last change a decade or so ago?
What's the documentation like? Are there any vignettes?
Does anyone else you know use the package?
4. Context matters
Also, where will you be using the package?
If I'm messing about on a toy project, then my criteria is much less than a mission
critical Shiny application.
5. Task Views
Something that people often forget about are CRAN task views: these are human curated lists of packages. Obviously there are no guarantees, but if you find yourself starting with a new topic, then it's worth having a quick look.
6. Let's get to work
In the following exercises, we are going to make these ideas a bit more concrete. Let's go!