New web-based tool designed to connect Pitt research faculty with collaborators, projects

By SHANNON O. WELLS

Given the often-daunting nature of digital communications, particularly within academic circles, a newly unveiled, Pitt-centric digital tool that uses research keywords to help faculty find cross-disciplinary collaborators and their projects hits the utilitarian sweet spot for researchers like Rob Rutenbar.

“There is not a week that goes by where I don’t get a couple of emails that go like this, ‘Hi, Rob, you don’t know me, but (this is who I am and what I’m doing),” he explained. “And I’m wondering if you know of anybody who is working on X. Because I am trying to put together a team, and I happen to know about Y and Z, but the proposal needs something about X. Who do you know?’”

Rutenbar, senior vice chancellor for research, is among those who welcomes the Research Discovery Tool (RDT), as a more effective, targeted way for faculty to connect with fellow researchers and their multifaceted projects.

“I know a lot of people at Pitt at this point, and I know people who know people at Pitt. And you know, sometimes the six degrees of separation thing works, but sometimes it doesn’t, right?” he said. “One of the things that’s fun about being at an institution as large as Pitt is that I’m always surprised to find that somebody was working on a thing I didn’t know they were working on.”

As Marcela Gomez, Pitt’s director of research analytics, presented it at the Senate Research Committee meeting on Jan. 17, the RDT could go a long way toward replacing such ambiguities with more precise knowledge of who is working on what, where they are and when they’re doing it.

She called the RDT an “advanced web-based tool that will allow the users to enter research keywords and obtain, as a result, a list of relevant projects by members of the Pitt research community.”

Development of the RDT, which Gomez said has been in the works for “several years,” started with a conversation with Rutenbar after she joined Pitt Research. As Gomez recalled, he asked her the simple question, “Can we help our faculty find collaborators?”

“(The RDT) is about advancing team science, and how it became a top priority of the University because it was also a priority of funding agencies across the board,” Gomez explained. “And of course, team science requires collaboration across disciplines to solve complex problems.

“Manually tracking all of the areas of expertise of our researchers is very challenging,” she added, “and we do not have readily available data to identify researchers in their areas of expertise.”

Before the COVID pandemic made virtual and hybrid work options common and work teams were smaller, Gomez noted, “it was easier to go down the hall and ask our colleagues, ‘What are you working on?’ and find collaborators. But as our different areas of expertise have grown, it has become more and more complicated to find out who can add value from a different perspective to our research endeavors.”

Meaningful searches

To create the tool, the Research development team partnered with Pitt Information Technology and Research Analytics, which manages the University’s research data. Amazon Web Services (AWS) was chosen to integrate data from agencies including the U.S. Patent and Trademark Office, National Institutes of Health and the National Science Foundation in a three-tiered access-control system.

“We are gathering data from the last 10 years, and we are using public application programming interfaces, or APIs, to gather this data and grab this data from the original sources,” Gomez explained. “In terms of general features … this is powered by the cloud services provided by Amazon.”

Gomez stressed that the tool is by, as well as for, the Pitt research faculty community.

“We wanted to make sure that the data we put in this tool is data focused on research that is happening at Pitt. We want to make sure that we are helping our researchers and our faculty find colleagues across the University,” she said. “Also, with the data that we have internally, we are able to identify these ambiguous names in a way that external tools are not able to.

“So we know that, for instance, ‘Rob Rutenbar’ that we find in our data is our senior vice chancellor for research, and is not another person and another university across the world.”

Partnering with Microsoft Azure Cognitive Services on a proof of concept enhances RDT’s utility and versatility, Gomez noted. “This will allow us to go beyond just looking for specific words, but add some meaningful search — or semantic search — into the equation,” she said. “So we built that proof of concept. It was very successful with the support from Microsoft, but then we needed to make sure that this can go live … We started identifying some roadblocks that would prevent us from bringing the RDT to production.”

This led the developers “back to the drawing board” with their Pitt IT partners. The parties created a University-based “data lake” using the AWS cloud system to organize files for efficient access.

“They are doing great work at modernizing our data warehouse, and we wanted to be part of that project,” Gomez said. “Because we want to make sure that this tool is something that is going to last and not have to be rebuilt in the process.”

Finding stuff, securely

With version 1.0 of the RDT complete, Research Analytics is now building a roadmap for “how we’re going to continue to maintain this, how we’re going to make it better, and of course, how we’re going to address the feedback that we get from our users,” Gomez explained. “And of course, we want to make sure that we continue our strong partnership with Pitt IT and Amazon Web Services.”

The RDT’s access page was designed to look “like every other search engine that we normally use,” she said, including a search bar and filters to narrow down results. The password-protected tool is accessible only to those with Pitt credentials. The latency, or time it takes for the process to generate results is among the aspects still being worked out, Gomez said.

Rutenbar said the RDT promises to be a notable improvement over standard search engines, including Google, which he said was “wonderful,” but not “perfect.”

“It misses stuff. It miscategorizes stuff,” he said. “One of the big, big things that we’ve done in this effort is … name disambiguation. If you type my last name into Google, you find me and you find my relatives, OK? You type (‘Gomez’) in, you find a gigantic universe of people who are not Marcela, right? … We are in there managing that stuff up front so you don’t have to."

Calling sites such as Google Scholar and Web of Science “OK,” they’re also “incomplete, unless you as a faculty member, are out there aggressively curating your stuff,” he added. “If you’re a STEM (science, technology, engineering, mathematics) person and you’re on top of what you’re doing, Google Scholar is pretty good.

“The further away you get from the STEM ecosystem, some parts of the clinical ecosystem are not well represented there. If you are on the creative arts side of campus (or humanities or social sciences) side of campus, or some of the professions, it doesn’t work so well.”

Responding to a question about why databases like NSA, NIH and NSF are available through the RDT search, but not the Pitt-based PERIS portal, Rutenbar said while NSF and NIH are public, everything in PERIS “is incredibly confidential.”

“Now, is it there? Yeah. But when you submit something into, pick your favorite funding agency, that is a highly proprietary, highly confidential thing … It’s OK for me to know about it, because, honestly, I own all of the data assets for all the proposals that are in progress … (But) I can’t open that stuff up to the public.

“We want to make sure we can find stuff,” he added. “We (also) want to make sure that when there’s a very reasonable faculty expectation of privacy, we don’t let people find stuff. We’re being rather conservative on how we put the locks in place. We’ll see how this goes."

Accessed at www.researchdiscovery.pitt.edu, or pi.tt/researchdiscovery, the ADT is scheduled for official release on Feb. 20. 

Shannon O. Wells is a writer for the University Times. Reach him at shannonw@pitt.edu.

 

Have a story idea or news to share? Share it with the University Times.

Follow the University Times on Twitter and Facebook.