Researchers and data scientists across the UW-Madison campus are developing and sharing Covid-19 models that can inform responses to the pandemic. A new service offered through the UW-Madison Division of Information Technology (DoIT) Cyberinfrastructure Group, which is available to all UW-Madison researchers, is helping these scientists collaborate and share their work.
Brian Yandell and Steve Goldstein, American Family Insurance Data Science Institute, and graduate students Srikanth Aravamuthan and Sean Kent are utilizing the new RStudio Team service. With data from an academic publication on Covid-19 screening strategies for college campuses, they used RStudio Team to create an interactive dashboard that helped campus leaders understand how changes in practices such as testing frequency and community behavior might impact decisions about reopening.
“Between Saturday and Wednesday, Sean and Sri were able to create the code behind the dashboard in GitHub and publish the dashboard on RStudio Team,” says Yandell. Within a week, the paper’s lead author, A. David Paltiel, shared this dashboard with a national audience. It has received thousands of views by users across the country.
While free, open-source versions of RStudio can run on an individual’s desktop, RStudio Team enables people to work together across collaborative environments. Because RStudio Team is hosted in the cloud on Amazon Web Services (AWS), it can provide more computational power than individuals have on their local machines. RStudio Team has two parts: an integrated development environment (RStudio Server Pro) and a visualization and publishing platform for data reports and graphics developed in R and Python (RStudio Connect).
“The big benefits of doing data science in the cloud are collaboration, sharing and scaling,” said Yandell. Using RStudio Team, researchers can share their work within research teams, with the broader campus community and publicly.
Another benefit of RStudio Team is that it allows collaborators to use popular Python and R interfaces in one place. “While R users are very familiar with RStudio Desktop, project teams are looking for more ways to work collaboratively, which is possible on this cloud platform. We’re also seeing a growing need to use both R and Python in the same project. RStudio Team allows users to intermix R and Python code in the same document,” says Tobin Magle, RStudio Service lead in DoIT’s Research Cyberinfrastructure group.
Kent, Aravamuthan and Goldstein have used RStudio Connect to develop other Covid-19 model visualizations using both R and Python, including coronavirus growth rates and infection transmissibility in Wisconsin. They have customized these tools for use by the Marshfield Clinic and Gundersen Health System to anticipate small, county-level Covid-19 outbreaks.
Kent appreciates how RStudio Team lets him easily share work with thousands of people, as his team did with their Covid-19 dashboard.
“Traditionally, researchers get the word out about their work by publishing a paper,” says Kent. “RStudio Connect allows people to also get their word out in a more interactive way. It’s really cool if you can work locally and a lot of people can access and get value from your work. I would definitely recommend this tool to other data scientists.”
DoIT is offering the RStudio Team service through April 2022. Usage will help DoIT determine if an ongoing RStudio Team service will enhance data science research and collaboration on campus.
To learn more about this Data Science Platform offering, visit the Research Cyberinfrastructure website. Additional information and webinars are available from RStudio. The Data Science Research Bazaar features two sessions about this platform: a poster and an upcoming workshop session on how to use this platform on Feb 24.