Researchers support college in making data-driven decisions
Behind the scenes of some big decisions, a student-faculty research team is sifting through extensive amounts of COVID-19 data to provide insight to campus leaders.
When COVID-19 caused a shut down of campus, Associate Professor of Chemistry Jai Shanata joined a group of campus leaders working on a process for the safe return of students. When he realized there would be an unmet need for understanding public health data, he turned to two of his research students.
“Very quickly we realized that if we want to make the best quality decisions for Cornell, we need to get the best quality data that we can get,” Jai said.
Senior Robyn Luchesi (biochemistry and molecular biology and philosophy double major) and sophomore Gwen Paule (biochemistry and molecular biology and Spanish double major), were ready to get started.
“Neither my students nor I had the slightest idea what we were getting ourselves into,” he said with a smile.
They needed to recruit one more person–lecturer in statistics at Cornell and Jai’s wife–Brandi Shanata. Brandi agreed to volunteer her time as a statistical consultant to help with the project.
The group now meets, virtually, about twice a day, and provides a lengthy weekly report to the crisis management team with their recommendations and data findings, so campus leaders can make informed, data-driven decisions about the next steps for the college community.
Cornell College is following a phased approach to resuming on-campus operations.
In order to move from one phase to the next, several internal and external conditions have to be met. Some examples of internal conditions include having sufficient supplies of hand sanitizer, preparing student quarantine spaces, and ensuring COVID-19 testing is available.
The external conditions portion of the phased plane is where the data research comes in. After some testing and a lot of literature review, the team settled on three domains to follow closely.
“First, we wanted something that tells us about the spread of the disease in the area. Second, we wanted to look at the use of resources–do we have the ability to handle if there is a surge in cases? Finally, the third domain is the severity of the disease. We want to know, for example, if the disease has mutated and is more severe or if we have better treatment or case management and the severity of the disease goes down,” Brandi said.
As one tiny piece of the data they are collecting to answer these questions, the team has learned the importance of using the positive test rate in their research. That’s the number of positive tests in relation to the total number of tests, instead of just using the count of positive cases. They have discovered that the number of tests performed every day across Linn and Johnson counties varies widely, making the positive test rate a more reliable metric
The team analyzes the data in their weekly report using green, yellow, and red zones to make recommendations about whether to proceed to another phase. This strategy empowers the college to move between phases in response to trends in the data, rather than swinging in and out of phases because of brief spikes or dips. Like a stoplight:
Green: data shows everything is okay to proceed forward.
Yellow: data shows Cornell should pause and wait to see what happens.
Red: data shows we should be concerned and consider moving backward in phases.
From the very beginning, Brandi advised that the group needed to take ownership of the data and start collecting the information that was important to Cornell’s decision-making process. Now, the students have a spreadsheet with around 30 tabs. Each night they add data from various public health websites in Linn and Johnson counties, as well as the Iowa Department of Public Health.
“If I could put the amount of data that they are sorting through, collecting, and organizing in terms of books, this would be an epic novel,” Brandi said. “This is not a novella, this is not a typical novel. This would be a Game of Thrones book.”
Despite the long nights spent deciphering data and writing reports, the students agree, it’s a rewarding process.
“It is amazing to work with two professors who I really look up to. Their passion and wisdom are invaluable,” Luchesi said. “Working on a project that has real-world applications, especially regarding public health, is something that is showing me a lot about what will possibly concern my career in the future. I am also lucky that I get to do something in a time where it is difficult to feel like you have an impact on something. I am glad I can do something to help, especially for my fellow students.”
Paule is also excited to work with the Shanatas on a timely project that will have a big impact on her life and that of her peers.
“This research opportunity is very unique due to these strange times and is extremely important to me,” Paule said.
This team’s main goal is to present the data in an unfiltered way so that Cornell College will know the reality of what’s happening close to home.
“We will be one of the smallest colleges to have such a large amount of customized data at our fingertips to make these very crucial decisions,” Brandi said.
They say that’s important, especially when there are so many messages about the disease and what’s happening with case-counts across the country.
“We want to present information in a useful and meaningful way so that the Task Force can make decisions about when we can safely repopulate campus and keep track of the area when we do,” Luchesi said.
Jai worked quickly when the college announced there wouldn’t be any in-person summer research on campus this year to provide as many opportunities for students as possible.
He took on a larger-than-expected group of research students because many were planning to work in labs or offices where internships and projects were canceled because of the pandemic. After all, Jai and Brandi were both once Cornell students who know first-hand the benefit of research and mentoring with faculty members.
Paule and Luchesi are now two of six research students working with Jai. Three others are spending their summer investigating COVID-19 testing methods.
“Summer research is always about mentoring students, helping them explore career possibilities, and develop transferable skills,” Jai said.
Jai and Brandi spend time discussing how students can talk about their research in job interviews and write about their skills on a job application. The team has also been working on developing protocols so others can step in and easily take over the data collection process. In addition, they’re learning how to write about and explain their research for different audiences.
“They have really learned so much,” Jai said.
Paule and Luchesi, who worked from their homes in St. Paul, Minnesota, and Chicago, Ill., respectively, earned a stipend for this research. They say each day comes filled with challenges to overcome.
“The most interesting and challenging part of this project is the day-to-day collection of data. It is fascinating to wait all day to update something at the same time every night,” Paule said. “Everything seems very slow at times when we’re collecting data with minimal change. Other times it seems very quick when we’re collecting data with quite a bit of change on a given day. These differences in my perception of time on this project lead to intrigue and impatience, all at once sometimes.”
This team has been following a systematic approach and tirelessly analyzing this data.
“Another challenging part of this project is that we are dealing with real data about real people,” Luchesi said. “When you look at this data it is easy to see numbers, but it’s important to remember what the number of COVID-19 patients on ventilators really means, and it’s nice to see when that number goes down.”
Even with all of the challenges of this research, Brandi said the students have never complained.
“Their work really has highlighted what a Cornell student can do,” Brandi said. “Neither of these students is a data analyst and neither is majoring or minoring in statistics. We didn’t expect this project to be this big. The idea was just, ‘hey, we know data is in a lot of places. How do we make this easily accessible to Cornell, so Cornell can make data-driven decisions?’ These students jumped in and adapted. It was amazing. They demonstrated so beautifully what a Cornell experience will do. They were able to very quickly learn and apply everything.”