We often get to speak with well-known Operations Research practitioners and professors who are dedicated to inspiring students and advancing the discipline. This time, we are excited to share experiences and insights from Prof. Laura Albert from University of Wisconsin-Madison. You might know Laura from her blog, Punk Rock OR, or her work on sports analytics. We interviewed her recently to get her take on how Operations Research is faring amid today’s uncertainty and what OR professionals can do to make a difference for social good.
People in the OR community know you as a prolific researcher and blogger. Can you tell us briefly about yourself and how you came to work in this field?
I was a bit of a latecomer to the field. During college, I took a course on optimization and figured out that I was very interested in how you can use mathematics to support interdependent decision-making. My PhD thesis focused on aviation security and for most of my career, I have been passionate about public sector applications. I have worked on OR applications in cybersecurity, infrastructure protection, disaster response, policing and opioid abuse, and more. In a sense, the common thread of my research has been helping public institutions reduce risk with limited resources.
Speaking of public sector applications, you participated in INFORMS’ Doing Good with Good OR committee. Can you share some of the most inspiring projects you recall from across the years that make use of OR to address social issues?
I think the projects I’ve appreciated most are those that require courage to pull together a toolkit around applications. Certain problems are well-studied by the community because we have existing models. Other times, we don’t, so it requires courage to come up with the mathematical abstraction and an application that improves outcomes. You have to gather data, talk with different stakeholders who will be involved in using the system, and more. Sometimes, the model you develop will be the first time a government agency uses OR. Seeing these kinds of projects is what I liked most about this committee.
One particular project I remember streamlined the process of getting welfare to recipients in the U.S. It required courage to find out what the model could look like. The project studied the amount of people coming in needing help with their benefits, and how to optimize the distribution of what they needed. It was implemented by the government agency at the time.
I’d like to turn to another pressing topic: COVID-19. You recently published a case study about pooled testing which you used for a course on probability models. Are you aware of places where this strategy is being used? What about other OR applications to address the pandemic?
Pooled testing offers a way to expand testing capacity up to four times higher. If you have limited testing resources, it’s extremely useful. I was able to get data from positivity tests throughout the U.S. in July. Of course, the situation in the United States has changed considerably since then. In July 2020, many states struggled to process COVID-19 tests quickly, with some states taking more than a week to process tests. We did this multifaceted case study in class (read more about it here). This was a fascinating study and an example of how mathematics can be used to support testing strategies.
“Instead of running a test for each person, laboratories could pool together tests from small groups of people and analyze them all at once. For the ones that come back positive, tests could be rerun one at a time with unused portions of the original samples, achieving the same results using fewer resources.” - read more on The New York Times
At the recent INFORMS Annual Meeting, I learned that Cornell University is using pooled tests for COVID response. They need to test almost 30,000 people regularly (students, plus faculty and staff). This is an enormous task. Their positivity rate is less than 1%. That’s enough to significantly expand their testing capacity with pooled testing. It's allowing them to test undergrads twice per week for COVID-19.
Speaking of OR applications, nearly all my research topics are affected by the pandemic in one way or another. I do a lot of work in emergency medical services. It turns out that many people don’t want an ambulance to take them to hospitals these days. So, medical services are moving towards treatment in place and telemedicine. I’ve been looking into this change in paradigm. It’s challenging to have a system that can deal with both emergencies and calls that can be scheduled.
Another topic I’m studying is policing and opioid usage. How can you effectively divert drug users to rehabilitation instead of the criminal justice system? With the pandemic, sadly overdoses have gone up. We are currently collecting some interesting data that can lead to a unique analysis of this problem, and hopefully lead institutions to make different decisions compared to before.
In aviation, travel is still down almost 80%. In the long run, we’re envisioning how people can go through checkpoint screening with less time and friction, while respecting social distancing. Facial recognition and biometrics will also play a role to keep things secure.
It’s an exciting time to do research, there’s a lot of applications out there.
I know your big passion is sports analytics. Do you know of any good use cases for OR in transitioning to “a new normal” for sports?
Sports is a typical field for OR applications. Until recently, you couldn’t impact big design decisions like the schedule and playoffs. In a typical year, changing these things would be too disruptive but now, it's prudent to revisit this. For example, in baseball, it’s not clear what the season will look like next year. How would you set up a tennis tournament today? How do you come up with a design system for football playoffs? What can we learn from the data and what does that tell us about sports? What’s the impact of having a home field advantage without fans as spectators? We’ve never been able to study this. There will be questions that will affect how we view officiating and questions surrounding decision making within games. I’m interested in what people find once we start to study this.
It sounds like there’s plenty of opportunities to make an impact with OR. But part of making an impact is also spreading the word about your research. That’s what I’d like to turn to a recent article you published with tips on “Engaging the Media.” What made you publish this article?
I started a draft of that article about two years ago, but I finished it during lockdown this spring. I was inspired to see many colleagues like myself appearing in the media. It was not just OR professors, by the way, but academics across the spectrum. The public was hungry for expertise across disciplines. I felt I had to finish the article to help my colleagues use their expertise to move conversations forward, in health and OR in general.
If you were to list your top 3 tips to OR researchers who want to get their story out there, what would they be?
First, repackage your research for public consumption. Takeaways for academics may not be the same for the public. List 1-3 takeaways and make sure you always come back to your points. Keep them short.
Second, try to take advantage of your existing resources to get your story out. For example, talk to your department chair or your university’s press office. You don’t have to do it all by yourself.
Third, practice: get used to repackaging your story. Practice with students or on social media, for instance, to hone your skills.
What do you consider to be the biggest challenges for the OR field in, say, the next 10 years?
I see a couple that are tied in with opportunities. One of them is building robust and resilient supply chains. I can’t remember national discussions involving supply chains as much I have seen these last few months. There are chances to expand the application of methods addressed by OR tools. If we can make that happen, that will open more doors for OR professionals.
Another challenge is transitioning to study new applications that have emerged with and without the pandemic. I have been in this field long enough to remember when health applications was an emerging trend with most OR work focused on logistics and transportation. Now, healthcare is a big field of study for us. It’s a major challenge to stay relevant and we're very relevant with healthcare. The question is, what’s next?
Finally, there’s data science. We all know that data doesn't do anything on its own. The point is to turn it into information, insight, and better decisions. Being able to lay claim to our piece of the data science discipline is key. OR has traditionally been model-centric, not data-centric, but data will play a more central role in OR. We should take care not to displace the model-centric approach. It’s an important part of our field and identity. It will be interesting to see data and model centric research that’s pushing the envelope of what we can achieve.
What advice would you give to students who are reading this blog and thinking of continuing their career in academia?
We are really good at studying the science of decisions and it’s an exciting time to do research. Sadly, we also have a global recession going on. Universities have advanced our understanding of the pandemic, and a lot of vaccines are being developed by academic research. Universities are a powerhouse for economic growth and innovation. So, we’re going to play an important role. Education itself has also seen some impressive advances. In a way, we’ve disrupted ourselves. I’m amazed of what university professors are doing in the classroom, so I would say there’s never been a better time to pursue this career.
Follow Laura on Twitter for more inspiration.