Using Machine Learning to Inform Decision Making

Collaboration was critical to the work between University leaders, researchers, and IT professionals during the COVID-19 pandemic as they leveraged machine data to make the best-informed decisions possible for the campus community.

Since the University of Arizona had already established a mature information security practice before COVID-19, UITS was able to quickly expand the Splunk monitoring system to assist with strategic big data needs.

Splunk analyzes time-based interactions, thus providing insight into human behavior, such as how faculty, staff, and students are using IT systems either remotely or on campus. Because protecting individual privacy is a critical element in data security, the ISO team worked closely with the University privacy governance council to anonymize the data sets.

Splunk gave UITS technicians the ability to look at population density that could be measured using campus wireless access points. This was particularly instrumental for the University’s campus reentry efforts and data-driven decision making.

When the University moved to remote locations in March 2020, UITS began monitoring its most commonly used systems to determine IT service usage across the globe. The data showed active users and performance metrics for D2L, Zoom, Virtual Private Network (VPN), remote Wi-Fi site connections, and many other important systems needed for remote work and learning. Through this identification and monitoring of systems, UITS was able to expand services where needed, such as increased VPN connections.

Expanding University Antibody Testing Across Arizona

The University of Arizona’s College of Medicine faculty researchers were among the first in the country to establish a reliable COVID-19 antibody test in spring 2020. UITS technicians were given a role in the Test-Trace-Treat initiative to provide technical support for expanding testing across the campus, county, and state.

The University of Arizona’s testing initiative started in coordination with the Pima County Health Department and was soon expanded to five medical clinics in Maricopa County. UITS teams also ramped up locally as they were called upon to set up 220 laptops, 120 label printers, and 40 bar code scanners for the campus antibody testing process.

In total, the project included 33 testing sites across Arizona. WNC staff traded off working 12- to 14-hour days, paired with a field technician every day, delivering and installing site hardware across Arizona. They logged over 2,000 miles in a week.

Making the testing initiative an immediate success included collaborative effort between the College of Medicine (Tucson), College of Nursing, and additional help from University Libraries and the College of Social & Behavioral Sciences.

“It was the perfect example of how things can really be efficient, effective, and successful when everyone gets together with a plan. We just coordinate and get it done.”

Cloud Contact Center Improves Remote Work Capabilities

As campus shifted to a fully remote workforce, there was suddenly a need to modify how departments handled inbound calls such as technical support, athletics ticket sales or registering for a COVID-19 antibody test. Historically, telecommunications technology required staff to be on campus to answer calls, which became challenge in the new remote environment.

The 24/7 Support Center, which operates a campus call center, had already been testing a new Amazon Connect platform for nearly a year and planned to offer it as a new service in 2020. The cloud-based call center system not only had the ability to support a fully remote workforce, but could also be expanded to accommodate others who needed to use it. The timeline was accelerated overnight to begin deploying Amazon Connect to campus units so their staff could receive calls remotely on their computer or cell phone throughout the day.

UITS telecommunications, network, service management, and workgroup and network consulting teams partnered to onboard 28 campus groups in the new call center solution. Key to to establishing these centers was speed and scale. UITS launched call centers for many campus groups in less than 48 hours, with the capability of taking calls from hundreds of thousands of callers.