The art of the online interface

By Colleen Walsh

Life will soon be easier for researchers eager to analyze information from the Harvard Dataverse by harnessing the power of Sid, Harvard’s version of the open-source Open OnDemand dashboard that connects scholars with supercomputers, thanks to the new LOOP application being developed at Harvard’s Institute for Quantitative Social Science (IQSS).

Len Wisniewski, senior director of research computing at IQSS, leads the Sid team that co-designed the LOOP application with the Dataverse team. He says the new software is aimed at helping researchers in a variety of fields—from political science to psychology to the physical sciences and beyond—to streamline their work.

“By simplifying the process of sharing files between Dataverse and research computing systems, researchers can then spend more time focusing on their research to produce better and faster research,” said Wisniewski, who noted that requiring scholars to take additional steps to access the data hampers the actual work. “Those extra steps can become a hurdle to the point where researchers just don't leverage the power of research data sharing on Dataverse.”

“We have researchers who tell us for example, ‘I am doing X, and now I want to do Y, but I have to learn a whole other process to access data to do Y,’” said Wisniewski. “So the goal when you develop this kind of product is to satisfy that kind of very basic requirement of what the user needs, and avoid unnecessary extra technical tasks.”

Saving researchers time is a key goal of the new LOOP application. In the past, those looking to study trends or evaluate research data hosted on the Harvard Dataverse repository would need to first download the data onto their computers, then upload it to the FASRC clusters in order to run their analysis. With their analysis complete, users would then need to download the results from Sid back to their computers, then upload the findings back into Dataverse. It was a cumbersome process requiring time, patience and awareness of both systems.

Now, with the help of this new app, that four-step progression will be streamlined by accessing the LOOP application via a drop-down menu from Sid. Scholars will be able to seamlessly move files back and forth between Dataverse and Sid to complete all functions directly from their laptops, phones, desktops, tablets, or possibly even their Teslas.

“That’s the idea, ubiquitous computing,” said Wisniewski. “You could do all your research work anytime, anywhere, even from your car while driving across the desert, as long as you have an internet connection.”

Wisniewski calls the single dashboard of Open OnDemand (Sid) “the holy grail of high-performance computing.” High-performance computing, which includes research computing, has traditionally required users to deeply understand the complex underlying subsystems that comprise supercomputers.

“Researchers will now have this single place where they do all their research, and when they're done they can just go home, instead of having to choreograph applications and data among multiple environments. It’s a much better system that will lead to better work for everyone.”

Currently in its beta phase, Wisniewski and his team are looking for volunteers to help test-drive LOOP. “We'll immediately benefit from more user experience research,” said Wisniewski, who is hoping to further assess the application’s usefulness and ease of use with early adopters and determine if there are any other features users would like incorporated into the new program.