When a public emergency such as an outbreak of smallpox erupts, good decision making can mean the difference between life and death for thousands. The right allocation of hospital beds, medicines, supplies and medical staff can affect how fast and how far a killer like smallpox can spread.
Ajay Vinze and Raghu Santanam, both information systems professors at the W. P. Carey School of Business, observed that in many ways, the planning and allocation problem faced by health officials is similar to sourcing and supply issues that businesses must solve. Vinze and Santanam applied business best practices to such emergency situations, utilizing an analysis that is similar to the way corporations streamline their supply chains. These two researchers are using computer models loaded with data about population distribution, location and capacity of hospitals, travel patterns, etc., as well as information about the infection and mortality rates of a disease.
Within this data “clone” of the real world, the researchers can simulate an outbreak and test various emergency responses: how many hospital beds, doctors and doses of antidote are needed and whether they’ll be at the right place at the right time. In one scenario, all of the smallpox vaccine in the country may be warehoused in a central depot and then distributed to affected areas after the fact. In another version, the CDC distributes much of its supply to cities across the nation in anticipation of an outbreak. Health officials will be able to enter these kinds of implementation decisions into the model and observe the potential impact on the community.
The result will be a tool that will help municipalities prepare for an emergency.
The catch? The inputs that Vinze and Santanam plug into their simulation engine are just numbers from a database: the population of the city and surrounding area, the number of hospital beds and doctors already in the area, the time to deliver vaccine from a central depot, etc. Outputs are also a set of numbers: casualties, medical resources used and predicted medical needs, etc. Vinze and Santanam realized that the numbers are helpful, but complex problems are not always best understood by examining spreadsheet grids.
“Bottom line, we have a lot of data and we end up having to make sense of this data in a variety of ways, mainly from the perspective of policymakers,” says Vinze.
Policymakers need to be able to conceive of this complex problem in a different way. A collaboration with the Decision Theater at Arizona State University, a high-tech visualization facility that specializes in making data come alive, is giving Vinze and Santanam the extra dimension they sought.
“Decision makers are faced with large amounts of data and the best we can do [without the Decision Theater] is show it to them in a tabular form or some flat, 2-D kind of environment,” Vinze added. “It becomes difficult for them to conceptualize the process part of what that data implies.”
Decisions in the round
The Decision Theater is built on the assumptions that humans are visual creatures and that seeing is believing — or at least understanding.
The centerpiece of the facility is a 260-degree faceted screen that can display panoramic computer graphics or video content. Although a wrap-around screen provides a unique viewing experience, the Decision Theater is not just an IMAX experience. The key to the Decision Theater is what’s behind the screen: an 80-node high performance computing cluster, dedicated high speed network connections with direct access to teraflop-capable computer resources, and a team of two dozen which includes experts in a wide range of disciplines — graphics, computer science, mathematics, public policy, educational psychology and geology.
Deirdre Hahn, the theater’s interim director and information systems affiliate faculty member, says that the front-end display capabilities and the back-end human and technological resources make for immersive presentations that go far beyond the static PowerPoint slides that policymakers usually consider. What may start out as a satellite photo of a city and individual architectural renderings gets the Hollywood treatment when the theater melds it all into an urban development plan. The city of Carlsbad, California recently worked with Decision Theater to evaluate plans for its downtown.
“When we take highly complex, disparate data and integrate it in a visual way, that helps decision makers sync up their mind’s eye they can make a more informed and collaborative decision because they’re all on the same page,” says Hahn.
It was such transformations that convinced Vinze and Santanam to collaborate with Hahn and the Decision Theater.
“The better your conceptualization of the decision problem, then hopefully, the better the decisions that you make. And what the environment that’s available in the Decision Theater allows you to do is to have this kind of immersive environment where you can bring your data to life in a way that leads decision makers to think about the problem more innovatively,” says Santanam.
Watching an emergency unfold
With the resources of the Decision Theater, Vinze and Santanam can offer policymakers — in this case, Maricopa County health officials responsible for the greater Phoenix area — a new tool to experience a disease outbreak before it happens. Without the theater’s resources, the two researchers can only show health officials data tables and graphs, typical presentation fare. On the other hand, a day with the Decision Theater allows decision makers to feel like they are truly — but safely — in the midst of the emergency while the technical team weaves in graphics and geodata to better illustrate the impact of an outbreak.
Santanam notes that while the actual data inputs and outputs are the same, the new format makes it easier for a policy planning team to “get it” faster even if they don’t have technical math backgrounds. He likens it to the experience of seeing a building’s square-footage layout and blueprints as opposed to “walking” around and through a virtual mock-up of the finished product.
“It allows [health officials] to simulate different kinds of emergencies and live through them on a more granular level you allow people to experiment with them upfront and, if you will, make decisions with a full deck as opposed to being in a reactionary mode when an emergency hits,” says Vinze.
Such simulations also even out the discrepancies that can occur as one person looks at numbers in one way while someone else, seeing the same data, interprets it differently. Looking at the same visual representation of those numbers often clears up confusion.
The business application
At the heart of their research on the emergency response to health disasters, Vinze and Santanam draw a parallel to business decision making. Businesses produce and store massive amounts of data, much of it fairly unstructured: a supermarket’s database of shoppers’ purchases, a phone company’s collection of customers’ calls or a logistics company’s thousands of daily deliveries.
Vinze says the issue for companies is how best to make sense of the data and extract value from it, because despite the size of the repositories, companies may still lack the right information to make decisions. In those situations, modeling and visualizing can help decision-makers draw meaningful predictions from existing data points. “What our research is showing is that sometimes the reason you simulate situations and model them is because you don’t have exactly the data you need to make that decision.”
The two researchers envision a new era in data analysis where visualization permits policymakers to see data points and trends take shape in a meaningful way before their eyes; where translating databases to an easy-to-understand theater environment is common; where decision makers who are not “numbers people” can play out “what if” scenarios in advance so as to be able to react better if those incidents come to pass. Then, if the unthinkable happens and Vinze and Santanam’s disease outbreak models are put to the test, the health officials will have already experienced the scenarios in the cool-headed atmosphere of the Decision Theater rather than in the emergency itself.
In short, even if it’s their first time in the hot zone, they will feel like they’ve seen it all before.
This project is funded by an IBM T. J. Watson faculty research award, a grant from the Maricopa County Department of Public Health (administering U.S. Department of Health and Human Services funding)and support from Arizona State University’s Science Engineering Research Visualization (SERV).
- There are parallels between organizing an effective response to a public health emergency and ensuring that a company’s supply chain functions properly.
- Organizations now produce and store mountains of data. While data mining can yield hidden insights, the arcane, spreadsheet-based nature of data analysis is inaccessible to decision makers who are not good at looking at the numbers and grasping what they mean.
- The Decision Theater at Arizona State University specializes in helping decision-makers visualize information stored in a database. The facility features a 260-degree screen, a powerful cluster of computers and dozens of experts from varied backgrounds.
- Analyzing a numbers-based problem visually allows policymakers to understand the issues quicker (humans are visual creatures) and enables them to be on the same page when making plans based on the data.