A team of statisticians, data analysts and researchers created and have maintained Ohio State’s COVID-19 dashboard since its inception early in the fall 2020 semester. Credit: Courtesy of Eric Mayberry, Dawn Larzelere, and Andrew Romero
For those invested in tracking the spread of the coronavirus at Ohio State, the university’s COVID-19 dashboard is likely a familiar sight.
The public-facing dashboard went online when students returned to campus in the fall, but when the country shut down at the start of the pandemic in March, the same team of statisticians, data analysts and researchers that now bring Ohio State daily COVID-19 updates created the data collection infrastructure in less than a week to report infection rate and virus toll to the university’s leaders and public health experts.
“To have a dashboard, to be very transparent about what’s going on, it almost in a way communicates back to the student and their family the level of effort on campus to contain and manage the virus,” Henry Zheng, associate vice president of strategic analytics with the Office of Academic Affairs, said. “We know that if things get out of hand, if we need to pull back, we know why.”
When campus shut down last spring, Zheng said Ohio State leaders needed a way to track and understand infection rates, hospitalizations and deaths due to COVID-19. Originally, data was compiled into a PowerPoint presentation and updated every night.
When the university announced its plan to reopen for the fall in June, it knew a daily PowerPoint wasn’t going to suffice, Zheng said. Shortly after, dashboards were assembled.
Zheng said one reason the dashboard is so effective is because of its multidisciplinary team of health and statistics experts, data analysts and epidemiologists who were involved in both the dashboard’s development and decision to reopen campus.
Eric Mayberry, director of data and analytics in the Office of the Chief Information Officer, created the first prototype for the dashboard with Leanne Stanley, consulting research statistician in the Office of Institutional Research and Planning.
Mayberry said they used data visualization software Tableau to create Ohio State’s COVID-19 dashboards — an internal-facing dashboard for university administrators and later a public-facing dashboard for the broader community.
The tools necessary to process and report the data for Ohio State’s dashboards were set up in three days, Mayberry said.
“I came from Nationwide Insurance, where a just small thing that they’d ask would have been a six-month effort,” Mayberry said. “Us being reactive enough to do that was amazing.”
For about six months straight, the dashboard team performed quality-control checks after collecting data each day until the current, automated version was created, Zheng said.
Morley Stone, senior vice president for research at Ohio State, acted as the final gatekeeper for the original dashboard, reviewing the information each time it was released, Zheng said.
“Most of the time he did not have questions, but if he asked questions, we knew something was not right and we’d go back and check,” Zheng said.
Some data is not as useful as others, Stanley said, so she was tasked with paring down the information to present the clearest picture possible. For example, she said they found it better to track the number of tests conducted versus the number of people who have been tested because students test more than once.
Andrew Romero, a data engineer for the Office of the Chief Information Officer, played a key role in sorting the raw data for the public and private dashboards, Stanley said.
Romero said data comes from many places at Ohio State — in-house COVID-19 testing labs, Vault testing labs, the Wexner Medical Center, Ohio State’s Compass Health app, isolation and quarantine housing, contact tracers and the Student Academic Services — as well as the state of Ohio and Franklin County. This information is used mostly for contact tracing students and monitoring outbreaks. He said 50,000 files are processed each day.
“The testing part of your record is one small piece, and then we go out and get all that other connected data from all over the university,” Mayberry said.
The public dashboard is just a fraction of the total data, and many different parts of the university need data that the public does not see, such as personal and operational information used for contact tracing and compliance purposes, Zheng said.
When it came time for campus to repopulate and a public-facing dashboard to go online, speed was more important than perfection, Stanley said. The dashboard has been reworked and redesigned multiple times since its launch in the fall.
Stanley said the dashboard has been changed to reflect input from the public and university officials over the past year.
The most recent update allows users to view data by semester, Stanley said, and data is available in an Excel sheet for people to examine it themselves.
Currently, the dashboard is completely automated, and the public dashboard has been viewed about 1.7 million times, Mayberry said.
Once data about positive test results is received, the information is given to contact tracers within one minute, Romero said. He said he believes his work with the dashboard has been successful in delivering data efficiently to help minimize the spread of the coronavirus.
“You don’t get the opportunity that often to literally help keep people healthy with data skills,” Romero said. “I think it’s just a really good example of the power of data analytics in general. A lot of this wouldn’t have been possible without the level of insight into this data that we’ve been able to have over the past six to eight months.”
