Thursday, April 24, 2014
Last updated: 3 days ago

UBC postdoc strikes deal to turn 2,000 academic journals into data mine

Florian Knorn/flickr

UBC has struck a groundbreaking deal with a scholarly journal publisher to allow text-mining, or allowing computer programs to find patterns across broad swaths of research data.

Text-mining helps researchers find patterns across the huge number of research papers published each year, which was previously difficult and tedious to do without the use of computers. For example, a program can be used to search separate pieces of research for related sets of words, allowing researchers to make connections that would not have been possible with previous research tools.

The deal was struck after Heather Piwowar, a UBC postgraduate zoology researcher, tweeted that she was frustrated that she couldn’t text-mine Elsevier’s journals. Elsevier publishes over 250,000 scholarly articles every year in 2,000 journals covering topics in the health, life and social sciences. Piwowar said that not being able to mine data from these publications was hindering her research.

Piwowar was then contacted by Alicia Wise, Elsevier’s director of universal access, and her negotiations led the publisher to allow UBC to text-mine Elsevier’s vast repository of content.

According to Piwowar, allowing researchers to look for trends across multiple studies can be immensely helpful. “[A] certain medicine might work on a certain disease, and no one had thought of that before, but the fact they have common gene patterns or common symptoms is something that might be discovered by this sort of text-mining.”

However, most journal publishers maintain strict control over their content, rarely allowing text-mining.

“Publishers are reluctant to open it up because they want to have control and they want to have money. And allowing other people to mine the articles they publish, they’re afraid they’ll lose both,” Piwowar said.

Wise pointed out that text-mining is a recently developed technique, and as such, Elsevier is still in the process of figuring out how to handle the practice.

Moreover, Wise said there is a lack of communication between the publishers and universities, and sometimes schools don’t know that they already have text-mining access to some publications.

Based on how much UBC pays for its scholarly journal subscriptions, Piwowar said she thought it’s only fair that UBC researchers should be granted free access to their data. UBC currently pays nearly $10 million for its electronic licences, said Allan Bell, director of library digital initiatives at UBC Library.

Piwowar, as well as many others in the academic community, believes that the government is currently paying twice for research, funding initial studies and then paying for access to the published results through publicly funded universities. In Piwowar’s opinion, this is an inefficient system.

But for the growing number of academics pushing for more open access, there are signs that the tide may be turning. A 29,000-signature petition is currently asking U.S. President Barack Obama to allow free access to all publicly funded research produced in the United States. “If the U.S. does this, it will put pressure on the Canadian government to pass similar legislation,” said Piwowar.

“If taxpayers fund research, which they do, then taxpayers ought to be able to read that research without paying again.”