Grappling with IoT Growth on Campus
Universities share how the expanding Internet of Things is impacting their institutions, from the sheer volume of data to staffing issues and the overall potential for students.
Smart campus efforts take advantage of the Internet of Things (IoT) to gain efficiencies from automation and analytics, often starting with building systems and energy usage.
But as IoT efforts expand to other parts of the university, including student success, some universities are experiencing growing pains working across disciplines to gather and analyze so much data. For instance, a few years into a smart campus initiative called the Integrated Controls and Analytics Program, Stanford University (CA) is finding data quality and data management are raising unforeseen challenges.
"Data management is the biggest obstacle we have right now," said Gerry Hamilton, Stanford's director of facilities energy management. "It all comes down to scalability and sustainability. We have found there is an exponential growth of effort that happens every time you deploy one more system."
Getting one building control system to integrate with one cloud analytics application takes a lot of work, because several people with different subject-matter expertise have to be involved — including IT staffers who make connections to servers and configure firewall rules, Hamilton explained. If you have 100 buildings, all with slightly different control systems and multiple analytics applications, every time you do these integrations, it adds to the layers of complexity. "From a labor standpoint, all this smart stuff we are putting in place to save us 80 percent on labor costs and make us 80 percent more efficient is doing just the opposite," he said.
At Stanford, the energy metering systems are used for billing individual colleges and forecasting projections for the on-campus generating plant. In order to create verified billing data, the university has created its own data correction tools for times when there are errors or gaps in the data flow or IP reliability issues. Yet one problem is that new stakeholders in the colleges are getting access to the raw data in one-minute timeframes. "What if someone tries to re-create that verified number from the raw data?" Hamilton asked. "They might say, 'When I add up these instantaneous values, I am not getting the same total you got.' That is an issue. If we want to do reliable forecasting, we need to have a clean set of historical data. That means we have to disregard the bad data or come up with some way to automatically backfill with reasonable approximations."
Hamilton said it all gets down to figuring out the business drivers. "Are we trying to be perfect for the sake of being perfect? There is some responsibility that comes with access to the data."
From Tech-Infused Stadium to the Classroom
IoT efforts at Arizona State University have seen a similar pattern of "exponential growth."
Last year, Campus Technology described how the institution was outfitting its Sun Devil Stadium with sensors connected to the WiFi and cellular network, to collect temperature, humidity and noise data for use by facilities staff. Sensors can identify if a faucet anywhere in the stadium is left running after a football game is over, for example, to help cut water usage. ASU also worked on providing information through a mobile app on the availability of parking and wait time estimations for concession lines and restrooms.
The tech-infused stadium was created as a test bed for a larger investigation of and investment in IoT technologies at ASU. "One thing we explored was what beacons can and can't do and what we can and can't get from the network," explained Chris Richardson, assistant vice president of IT development at ASU. "Some of that experimentation influenced our broader approach to using the network."
The university has piloted the installation of location-tracking beacons in classrooms to understand trends in student attendance in big classes, where taking attendance manually is impossible. "We have turned a few of our classrooms into a lab to study how to check students in and out without much intervention," Richardson said. Because the provost's office has expressed interested in attendance, ASU wanted to see if it could collect that data in a way that doesn't expose students to unnecessary privacy risk.
"We decided to test out these beacons and have students' phones pick up what is happening," he said. ASU built an API to the student information system to pull in the record of the student, and with student and instructor permission they outfitted a few classes with the beacons. Interested students downloaded an engagement app on their phones. The app allows students to see if the room has a beacon. Once they click on it, any other time they come into the room, the beacon automatically picks them up. "Many students were interested in having their attendance in the palm of their hand, to know how they were engaged with classes," Richardson said. "In fact, we found that some students would use the app to check in manually in classrooms that aren't beacon-enabled."
ASU didn't share individual student data with the instructors, but it is using the pilot to rethink how it communicates with students via mobile app. "We have partnerships with groups that run student success in our provost's office, and we are building communication plans to engage students through push notifications," Richardson said. "We will be able to better understand why they may take action or not and determine if that is something we might want to impact with adviser relationships," he added. "It's not easy to see the path from the stadium pilot to the classroom to now a broader mobile app strategy, but they all build on each other."
ASU is not alone in trying to apply IoT to student success. A University of Arizona professor is using data collected when students swipe their ID cards on campus to see if it reveals patterns about student routines that could be used to predict their likelihood of returning to campus after their freshman year.
A news story on the UA website noted that Sudha Ram, a professor of management information systems, gathered and analyzed data on freshman student ID card usage over a three-year period. She then used that data to create large networks mapping which students interacted with one another and how often.
"Considered together with demographic information and other predictive measures of freshman retention, an analysis of students' social interactions and routines was able to accurately predict 85 to 90 percent of the freshmen who would not return for a second year at the UA," the story noted, "with those having less-established routines and fewer social interactions most at-risk for leaving."
Impact on Staff
Regardless of whether the goal is impacting student life or just building or parking lot efficiency, deployment of IoT systems requires building new relationships and new governance structures. Chuck Benson, assistant director for IT in Facilities Services at the University of Washington, said universities hoping to expand their usage of IoT systems must come to grips with the cultural differences between operational technology (OT) and IT departments. Facilities managers come up through trades such as electrical or plumbing. Culturally they are very different than central IT employees, yet all of a sudden they are coming together on projects, Benson said. "Facilities management groups are getting large, complex IT systems thrown in their laps, and they don't have experience with them," he said.
The IoT energy management effort at UW has required a team effort by a conservation manager, a mechanical engineer, an IT exec, an electrical engineer, a vendor and a subcontractor, who meet every two weeks to address meter management and data flow. "We have gotten to be a pretty good team," Benson said, "but three years ago that wasn't the case. We had lots of issues. But we all acknowledged we come from different places and kept hammering it out."
Hamilton said one lesson Stanford has learned in this area is that it is important to have people who understand both OT and IT. "For a while things were falling in that gap," he said. Finding people who knew both the OT systems and IT systems well was a major accomplishment for Stanford, starting about two years ago, he added. "That is a big problem, and unless most universities address that, projects are going to stall out."
Benson co-chairs UW's IoT risk mitigation task force to develop some governance and oversight of these systems. "They span so many organizations that ownership is not clear or nonexistent," he said. "With the task force, we try to bring groups together and tell stories from their perspective and we find that people do find common issues." For example, the person who runs campus networks and the person who handles energy management for the university see the same kinds of issues but from different perspectives. "You start to develop common language around it," he said, "and then you can get to risk mitigation."
ASU's Richardson admitted that getting diverse campus groups to come together on smart campus projects has been a challenge, but he said it is definitely getting better. "My advice is to start small and mutually come to an understanding of what is possible and prove what is possible with limited risk, and then iterate on that."
People may look for traditional payback or return on investment from these projects, and it is often not that simple, Richardson said. "It involves a complex set of relationships. Organizational silos can get in the way. When we talk about governance, it is a massive undertaking, but really exciting. It will take very wide involvement to make it work."