Skip to content

Latest commit

 

History

History
85 lines (61 loc) · 2.61 KB

general-recommendations.md

File metadata and controls

85 lines (61 loc) · 2.61 KB
Raw
title subtitle author date
Preparing R code for reproducible research
General recommendations
José R. Ferrer-Paris
March 2023

Follow suggested best practices

First take a look at suggested best practices and style guides for coding in R, for example:

It takes time to get used to all of these, don't need to try to implement everything at once.

General recommendations

Here are some personal suggestions that will help you in your journey toward reproducible code.

Know your tools

Input data

  • Share your (raw) data (at least a sample dataset)
  • Document data preparation and cleaning
  • Foolproof data ingestion:
    • universal, simple formats if possible (e.g. .csv, .txt)
    • use open access tools and libraries for more complex formats whenever possible (e.g. sqlite, gdal)
    • follow data from downloading to final data.frame/tibble

More on this: Making your data and code open access / CES Skills Seminars 20-04-2023

Documentation

  • README or DESCRIPTION file
  • Comment your code:

    • Comments in R script,

    • Documents or reports using R/Sweave and knit, Rmarkdown, Quarto, etc.

  • Vignettes and tutorials

Modular code

  • Short R scripts with explicit inputs and outputs
  • Save intermediate steps (Rdata or RDS files)
  • Run Rscript or Background Jobs
  • Use and reuse your functions !
  • Write a package !!

Output data is useful for other people

  • Presentation of your data outputs
    (i.e. images or pdf documents)

    • Useful for readers

    • Reviews

  • Fully automated reports (Sweave and Knitr!)
  • Outputs and raw results

    • Might be useful for meta-analysis

    • Tabular formats

    • Structured text formats (Metadata, XML)

    • Spatial data (raster, vector data)

More on this: Making your data and code open access / CES Skills Seminars 20-04-2023

Document your journey

  • Tutorials / blogs
  • Keep a change log: explain what and why (fixed errors, etc)
  • Version control (e.g. git and GitHub!!)