README.md

In maskedforreview/gtdnews: Online appendix for @@ paper

Note for reviewers

This is the reviewer version for the online appendix.

As described below, the appendix consists of three parts. The third part is most relevant for the review, because it contains the main innovations of this study. Also, this part does not require copyright sensitive information, so the data for this part is included in the GitHub repository. If you install the helper package as instructed below, and follow the instructions in the third part of the online appendix, the results presented in the validation and analysis sections can be replicated.

About this repository

This repository serves as an online appendix for [@paper]. It contains the code and data to replicate the analysis.

The appendix is composed of three parts.

• The first part describes how we download the GTD and The Guardian data, and prepare the Document Term Matrices (DTMs).
• The second part contains the preprocessing of the DTMs and the calculation of the similarity scores.
• The third part describes the validation and analysis.

If you are only interested in replicating the analysis you can skip straight to the third part, for which all required data is directly included in the package. To replicate the study from scratch, you will need to start at the first part, because we cannot include this data in the package due to copyright restrictions. If you are interested in applying our method for other data, you can start from the second part using your own DTMs.

Appendix helper package

This online appendix contains quite a lot of code, and requires several packages We have created a helper package that contains the code in documented functions, and has the required packages as dependencies. The easiest way to install this package (without cloning the repository) is to use the remotes package, that allows you to directly install it from github.

library(remotes)
install_github('maskedforreview/gtdnews')

maskedforreview/gtdnews documentation built on April 12, 2021, 11:53 a.m.