author_name
current_date
This is my current data-analysis project template.
I use a bash script that sets up a new project for me based on this template.
Otherwise, simply download this repository and extract into your projects’ directory, change the directory name and start editing files.
The project structure is as follows:
data/: Original data files, as they were received.
I don’t touch them at all. I may even set read-only permissions on them. I typically track versions (even if sometimes they are binary files) for reproducibility. Mind about confidentiality issues if the repository is to be made public.
doc/: Supporting documents, papers, etc. that are not produced by you, but are useful or needed for reference.
Binary documents in PDF or doc(x) formats are untracked (edit doc/.gitignore as needed). However there is a Notes.md for keeping a project log and a Makefile to compile the notes to PDF format.
public/: Dissemination material in HTML format (reports, slides). Everything here will be published automatically on line by GitLab Pages.
reports/: Analysis reports or slides in PDF format. Different from public/ in purpose as well. public/ is for dissemination and collaboration on line, whereas reports/ is for PDF documents, targeted to others or internal.
src/: Source code for everything. Functions, figures, reports, analyses.
The .Rmd file produces a RMarkdown report (for public/ or reports/ or both). You can make as many of these as you wish.
Write functions into functions.R. The rest are ancilliary files. Write macros into before_body.tex and specific packages into preamble.tex. Include bibliographic references in BibTeX format into the .bib file. You can change the bibliographic style in the .csl file by whatever style you prefer.
The .Rproj file defines a RStudio project with some preferences and need not be touched.
.gitignore specifies which files are not to be tracked.
.gitlab-ci.yml configures GitLab’s continuous integration in order to produce GitLab Pages based from README.md and files within public/.
README.md is the current file and should be edited to give an overview your project. It will become the index page of the GitLab Pages website, and can include links to compiled reports for dissemination.
Sometimes we want to display some results or run some code to produce the index page. In that case, use usethis::use_readme.rmd().
_targets.R is the main file of the project. It consists of a list of targets that need to be computed using functions and code therein. These targets are R objects than can be referred to from within the Rmd files to produce the reports.
Some of the targets load or clean up the data sets producing intermediate objects that are used by others in turn, others make figures and others are meant to compile the reports themselves.