README.Rmd

---
title: "README"
author: "Arthur Gailes"
date: "10/14/2020"
output: github_document
---

# Default Data Science Project Directory

The folders in this project are:

* _data_ - all the data you have collected or been given to analyze.
  + _raw_ - data from external sources, not to be edited.
  + _tidy_ - organized, munged, prunded, and/or joined data for analysis. I usually add this to .gitignore, since it should all be reproducable by the package code.
* _analysis_ - scripts for tests, internal reports, exploratory analysis and graphs. Naming convention is a number (for ordering), the creator's initials, and a short "-" delimited description, e.g. "1.0-jqp-initial-data-exploration". 
* _references_ - data dictionaries, manuals, and all other explanatory materials.
* _products_ - Anything that is published or distributed
  + _figures_ - publicized graphs, charts, etc
  + _reports_ - written reports (.docx, .pdf, etc)
  + _data_ - publicized datasets (.csv, .xlsx, etc)
* _README_ - project root documentation
* _.gitignore_ - instructions for git to ignore files. 
* _R_ - R package functions, if necessary (see below)
* _src_ - python scripts and functions

Obviously all folder structure is flexible; this is only a starting point to make sure our analyses orbit around a different structure.

To set up an R package directory, run this code in README.Rmd or in the console (making sure you're in the correct directory)
```{r eval=FALSE}
library(here)
knitr::opts_knit$set(root.dir = here())
# usethis::create_tidy_package(path=getwd(), "your_package_name")
```

Note: all .gitkeep files can be deleted; they're only here so the folders will show up.

Based on:

* https://peerj.com/preprints/3192.pdf
* https://nicercode.github.io/blog/2013-04-05-projects/
* https://drivendata.github.io/cookiecutter-data-science/#directory-structure