Sci-Hub And Its Users | In the Pipeline

"I’ve been meaning to link to this article by John Bohannon on Sci-Hub, the gigantic paper-sharing (well, paper-pirating) service. The person behind the site (Alexandra Elbakyan) provided him with a large amount of data on the download requests it receives, and the overall picture is very interesting. Nation by nation, the single biggest user of the site is Iran, followed by China, then India, then Russia, then the US. The Iranian figures are apparently inflated because of massive local mirror sites in Tehran and other cities, but that also means that there’s a lot of traffic that future estimates will miss. On a population-adjusted basis, the big surprise for me was Tunisia, who are in the top ten. (Morocco and Egypt make strong showings, too). If you break down the requests by publisher, it’s Elsevier in a landslide. And when you look further into the site’s usage in the wealthier countries, you find the requests cluster around universities, where many people presumably already have access to many or most of the papers that are on the site anyway. An interesting note is that the top US city resolves to Ashburn, VA, which could be the Howard Hughes Janelia campus, or could be the servers from the Wikimedia Foundation, but also seems likely to be a huge amount of use from George Washington University, whose library system has been vocal about the amounts its being charged by scientific publishers. One thing that everyone would be interested in is Sci-Hub download traffic versus that of the actual publishers, but here it’s the publishers that are keeping things under wraps: no one has the figures. An Elsevier estimate from a few years ago makes it seem as if Sci-Hub is still only getting about 5% of the above-board publishing traffic, but it’s really impossible to be sure. As for total number of users, it’s well up into the millions, especially when you consider shared IP addresses around universities and the like ..."