In addition to providing a curated catalogue of data sets and analytic tools tailored to environment-security research, one of the primary goals of the DANTE Project is to present use cases that set out to replicate methodologies employed by modern high impact peer reviewed studies in human geography, political science, and environmental science. While developing the replication studies we will provide several tools designed to make advanced analytical techniques more accessible to begginer and intermediate users.
The ability to parse leading peer reviewed research, comprehend their underlying methodologies, and adapt them to personal needs is a major barrier to early-career scientists with high technical aspirations. This already difficult task is made even more challenging by publications with inadequate written methods plagued by poorly conveyed data processing procedures, insufficient details regarding software and package use, and a lack of specific arguments used for statistical modeling. All these issues would be remedied if researchers were required to provide their underlying code along with submissions. Not only would this provide total transparency, but it would serve as a teaching tool for graduate students and early career scientists.
Sadly, replication is one of the most overlooked aspects of the scientific process. Only recently have journals began to require underlying data used in analysis be submitted alongisde the manuscript. Even fewer journals require code used to generate analysis be submitted. This is especially problematic in environment-security research where the signals are notoriously weak, and policy decisions based on research may have lasting global impacts. We developed
duplicator with the goal of promoting reproducible open-access research. The replication studies developed by the DANTE platform are contained entirely within the
duplicator package. All data sets, analysis, and written documents can be reproduced using the functions, data, and vignettes embedded within the package. Moreover, all data acquisition and pre-processing steps can be re-produced using our raw data scripts. The end result is a completely transparent and open-source research and teaching platform.
You can install the current version of
duplicater from GitLab by running the following command in your R console:
devtools::install_gitlab("dante-sttr/duplicator", dependencies = TRUE, build_vignettes = TRUE)
To install R packages over Git on a Windows system, you must install Rtools first. The latest version of Rtools is available here. Furthermore, you may experience difficulty installing R packages over Git if you utilize a Windows machine on a network with Active Directory or shared network drives. To enable proper package installation under these circumstances please follow this guide.
A full list of available functions developed for
duplicator replication studies can be accessed through the online reference manual. The underlying code for functions developed for use in
duplicator replication studies can be found in the GitLab repository here.
Nearly all data sets used for
duplicator replication studies are available as exported objects and are fully documented with title, description, minimal metadata, url for hosted location, and official citation when available. To retrieve dataset documentation, call
help() on the dataset in question, or browse the full list of available data sets in the
duplicator reference manual. Raw code used to produce embedded datasets can be accessed here.