Thread by @cbirunda

Sharing @JennyBryan's deep thoughts (aka pro tips) from "What they forgot to teach you about R" workshop at #rstudioconf.

1. Be organized.
- Future you and others will thank you!
- Make sure the projects can explain themselves by using “wordy” and “needs” words
- Set up a clear structure for the sub-folders: (e.g. data, fig, syntax)
- README is great but it depends on how well it is maintained

2. Adopt a project oriented workflow
- You can work on more than one things at time
- You can easily share, reproduce and collaborate.
- Having a separate RProject for each project will keep everything contained and again organized!

3. Practice “Safe paths”
- Don’t use file paths as strings. (e.g using paste or strsplit to combine file paths)
Instead install fs (file path handling) and here packages.
- The here package is designed to work inside a project. It helps build paths within a RProject.

3. Practice "Safe Paths" (Continued)
- This is particularly useful if you work in team...each team member does not have to manually edit file paths.

4. Names matter! (1/2)
Files names matter. It must be
- machine readable: no white space, punctuation, accented characters, case sensitivity, generally exclude any weird characters
- human readable: easy to understand, name contains info on file content and context

4. Names matter! (2/2)
- Lastly, it must sort nicely: computer automatically sorts files names - don’t fight against its nature.

5. Break logic and output separately. (1/2)
- Pulling data processing steps apart so it’s clean and easy to follow-up.
- Create separate files for different processes: reading_data, wrangling_data, modelling, make_figs, etc
- Sketch out the purpose of each script (see pictures)

5. Break logic and output separately. (2/2)
- Create a final and separate controller script that runs all of the files (e.g. like dominos)

6. Important: always restart your R session between scripts!! You should self-contained R scripts for a smooth ride forward.