This compendium is for a research project in progress, with the aim of contrasting information from fake Twitter storms with real hurricanes, using the case of Hurricane Dorian in the United States.
This research was developed by Joseph Holler.
This repository is based on a template that contains a folder structure, template documents, and suggestions for best practices for conducting reproducible geographic research. The folder structure presented here can be used to:
- pre-register, document, and share original research, or
- document, organize, and share a reproduction and/or replication of original research.
An overview of the folder structure of this repository is provided below. The README.md
and index.csv
files contained in each folder provide details about the structure of that folder and suggestions on its use. The docs/report/
folder contains templates to facilitate 1) the pre-registration of original research, and 2) report the complete details of attempted reproductions and replications of original research.
Reproduction of <Study Name>
|- docs/ # study documentation
| +- report/ # reproduction plan, reproduction report
| +- manuscript/ # manuscript components
|
|- data # raw data, which are not changed once created
| +- raw/ # raw data, will not be altered
| +- derived/ # derived data, will not be altered once created
| +- scratch/ # temporary files that can be safely deleted or lost
| +- metadata/ # documentation of metadata
|
|-procedure
| +- code/ # any programmatic code, clearly named and commented
| +- protocols/ # any non-computational protocols
|
|- results # all output from workflows and analyses
| +- figures/ # graphs, likely designated for manuscript
| +- maps/ # maps, likely designated for manuscript
| +- other/ # diagrams, images, and other non-graph graphics
|
|- readme.md # the top level description of content
|- study.Rmd # executable Rmarkdown , if applicable
|- Makefile # executable Makefile, if applicable
|- study.Rproj # RStudio project, if applicable
Every research project is different. This repository is designed to serve as a flexible guide capable of structuring work completed throughout the lifecycle of different types of research project. No matter the project type, a few key suggested practices when using this repository include:
- Keeping original, raw data in the
data/raw
folder. Do not alter that file during data analysis. - Keeping data derived from the raw data (e.g. subsets) separate from the raw data in the
data/derived
folder. - Keeping Exploratory/experimental outputs in the
data/Scratch
folder. Files in this folder should be able to be deleted without negatively impacting the project. - Limiting manual changes to data. Conduct as much data processing and analysis as possible with code.
- Keeping code in simple text files that are human readable.
- Explicitly tying procedural documents or code together so it is clear how things work together
- Creating a top-level
Makefile
or Rmarkdown file that documents computational work in executable form. - Providing a executable electronic research compendia that shares the entire computational environment
- Keeping a README in each folder, describing the purpose of the directory and its contents.
- Maintaining a formal metadata descriptor at the root of the package that describes all the important input and output data files.
The structure of this repository closely follows the excellent rr-init repository, which in turn follows Nobel (2009). We have also incorporated structural ideas from Gandrud (2015) and Camerer et al. (2016, 2018). Reference information related to the Pre-registration Template and Replication Report Template is included in the /docs
folder.
This research compendium template is in its testing phase. It will eventually be published to the public with an open-source license at https://github.com/HEGSRR/RR-Template The compendium is the work of Peter Kedron, Joseph Holler, and Zach Hilgendorf. This draft version of the compendium is provided for private use only: not for distribution.