By developing a forecasting model and interactive dashboard, two groups of USC Viterbi students are helping Keck predict patient demand — and plan accordingly. (Image / ChatGPT)
Operating a medical center is an incredibly complex task. Medical professionals have to juggle numerous patients and ensure high-quality care while considering limited resources like time, staffing, and tools.
With this crucial yet difficult priority in mind, the team at the Keck School of Medicine of USC turned to data analytics and forecasting to better anticipate future patient demand and plan hospital resources. Maybe, this delicate balance of quality care and limited resources could be assisted by an analytics system that predicts the number of inpatient admissions. A forecast like this could help them decide, for example, how many nurses they might need to have on a shift one month from now.
Two groups of master’s students at the USC Viterbi School of Engineering set out to help the Keck team tackle that forecasting challenge with an eye toward more efficient operations. By developing a forecasting model and interactive dashboard, the Viterbi teams are helping Keck better understand the trends and patterns within its own medical operations. That, in turn, will help Keck predict demand on its facilities and resources — and plan accordingly.
“They were very well-prepared and so excited to work with us,” said Miranda Li, the director of the office of performance and transformation at Keck Medicine of USC. “I was very impressed. They were more professional than a typical student, and asked a lot of good questions.”
This project is still in development, and will wrap up later this summer. In order to complete it, the team Viterbi has put together came in two parts.
First, students in the health systems management engineering master’s program at the Daniel J. Epstein Department of Industrial & Systems Engineering (ISE) were assembled by David Belson, an adjunct professor at USC Viterbi. This team handled the operations side first, understanding Keck’s needs and priorities to implement that into the final model.
“They needed to know what to expect, in terms of how many surgeries they would have or how many patients would be admitted, for example.” Belson said. “Managing all those resources at a big hospital is a real challenge, and they were hoping they could be able to plan better by using forecasting models.”
When it came time to create the forecast itself, students in the analytics master’s program at ISE were brought on, supervised by Bruce Wilcox, an associate professor of industrial and systems engineering practice at USC Viterbi. These students are in charge of building the forecasting models, and now, they are creating a Keck-friendly dashboard on the business intelligence platform Tableau.
Belson has been in the hospital management industry for decades, and he understands just how important managing one’s resources can be for a medical institution like Keck.
“Budgeting is pretty fundamental,” Belson said. “If the hospital has too many staff or too few staff or does not anticipate demand, it’s pretty important. Because of that, this tool should become very helpful for Keck.”
This tool could become baked into the day-to-day operations of Keck, helping them adapt to any scenario while taking operations and budgeting concerns into consideration.
“They can plan staffing decisions ahead of time. If the patient demand will be very high, they can increase nurses accordingly,” said Iris Xia, who led Wilcox’s data-focused team on their mission to refine the data behind Keck’s operations. “I was very excited about this, because we can really influence their operational workflow – our end goal of this project is to help their operations.”
Despite the weight of the project, the Viterbi students were more than up for the challenge – in fact, they led the charge throughout development. Wilcox emphasized the students’ interactions with the team at Keck in weekly meetings, encouraging them to hone their skills not only in engineering, but also in project management and consulting.
“I have zero direct interaction with these clients that don’t involve the students, I really want the students to do all the interaction with them, and I think that’s part of the learning experience,” Wilcox said.
One of those standouts is Xia, who was new to this management role. She approached the organization of this project democratically so that her entire team could enjoy the project’s development.
“I just took the responsibility to manage the team and the communication with the client,” Xia said. “We’re developing the forecasting and Tableau dashboard at the same time, and so I asked my team members if they had any task preferences, so I could assign tasks accordingly.”
In practice, the Keck team is hoping that this forecasting model and Tableau dashboard will reshape the way they prepare for an influx of patients. Whether it’s increasing staffing, anticipating challenging upticks in patients, or managing their sprawling facilities, this forecasting model will offer a better window into how they can properly manage their operations.
“We weren’t forecasting how many patients we were going to admit, we did more small-scale forecasts – and we never formalized the process,” Li said. “This model could really help us plan out resources, and that will also give our leadership flexibility. They could look at any scenario and see the resources required to handle that future situation.”
Published on August 17th, 2026
Last updated on August 18th, 2026