Site icon IT World Canada

N.C. State finds research partners using data analytics

North Carolina State University has signed up IBM Corp. to help its technology transfer office speed up the process of matching university research projects with potential investors and industry partners.

Under the effort, the university will use IBM’s Apache Hadoop-powered Big Data analytics technologies to sort through massive volumes of Internet data to find potential matches.

Two pilot projects that were designed to test the IBM technology have already proved successful. One test earlier this year involved finding potential investors for a new Salmonella vaccine being studied by university researchers.

The analytics software scoured 1.4 million Web pages spread across corporate sites, blogs and social networks and returned a list of potential matches in less than a week. By contrast, the university’s seven-member technology transfer team would have needed two to four months to do the same thing.

The success of that pilot project, and of another one involving a targeted drug-delivery system, persuaded the university to apply IBM’s data analytics technology to such searches in future.

Much of the need for automation stems from the amount of work, said Billy Houghteling, director of the Office of Technology Transfer at N.C. State. “Resources are often a challenge. We have a large and diversified IP portfolio,” he said.

At any time, close to 3,000 research projects are under way, Houghteling said. Finding potential investors or partners for the projects is handled by a seven-member team at the office and by about 12 additional support staff. On average, the group spends three to four months on ‘triage” just to identify suitable industry partners, he said.

It’s a task that involves combing through countless Web sites, company documents and financial statements to figure out whether a company might be worth approaching for help, he said.

“Now we can surf millions of pages of information and have as an output a prioritized list of partners” ranked by the likelihood of their willingness to invest in a university research project, he said. “We can very quickly identify 10 to 15 potential partners.”

During the pilot tests, the list generated by the IBM analytics engine was not only similar to the one compiled manually by the technology transfer team, it also included potential partners the team had missed.

The data analytics capabilities will be delivered by IBM BigSheet, a relatively new analytics engine that is designed to extract, annotate and analyze large amounts of unstructured information The university will also use IBM LanguageWare text analytics tools and the company’s Cognos Content Analytics tools for analyzing unstructured data. All three technologies run on IBM’s distribution of Apache Hadoop open source software.

The N.C. State project is a good example of how emerging smart analytics technologies can be applied to an existing problem, said Chris Spencer, an emerging technologies strategist with IBM’s Software Group. “This is an interesting Big Data problem” the university is attempting to resolve, he said.

Jaikumar Vijayan covers data security and privacy issues, financial services security and e-voting for Computerworld. Follow Jaikumar on Twitter at @jaivijayan or subscribe to Jaikumar’s RSS feed . His e-mail address is jvijayan@computerworld.com .

Exit mobile version