Opening Pandora's Box: The Content of Sci-Hub and its Usage

This archive contains all you need to re-run the analysis as performed in the corresponding manuscript.

The data.tar.gz contains both:

  1. The raw data, including the 62 million resolved Sci-Hub DOI
  2. The aggregated data used for the analysis
  3. The scripts used for resolving the DOI and to analyze the data

If you unzip it you'll find that it's all packaged up in 2 main folders:

  • data/
    • data/raw/ contains both the complete corpus as well as all downloads
    • data/aggregates contains the aggregated data & analysis results
      • *data/aggregates/aggregate_downloads** contains the data for the download set
      • data/aggregates/aggregate_finished_binomial.csv contains the data after the binomial test for over/underrepresentation.
      • over/underrepresented_publishers.csv only contains the ones being over/underrepresented with a significant, FDR-corrected p-value.
  • figure_scripts contains the scripts for plotting the data as well as performing the binomial test (figure3). In addition it includes the Ruby script for resolving the DOI (doi_resolve.rb).

The data subfolders contain a README to explain the columns of each data file.

The raw data resources before resolving the DOI come from

  • DOI:10.5061/dryad.q447c (Downloads)
  • DOI:10.6084/m9.figshare.4765477.v1 (complete corpus)

All the best, Bastian