Show Notes
Mandy Hering never expected her career to flow into water research. But today, her statistical science expertise addresses global needs and supports municipal water systems across the country in pursuit of new ways to deliver clean water. Discover how the Baylor mission and statistics intersect to protect water resources for the future.
The conversation highlights:
- The versatility of statistics and how Baylor researchers collaborate across fields ranging from biology and environmental science to address human need.
- The need Hering recognized for statistical scientists to specialize in water needs ranging from acute municipal challenges to broad global trends
- Practical ways Hering helps cities forecast and manage water quality, saving energy and improving infrastructure.
- The evolution of data science and working to lead in an ever-changing field.
- Her passion for training the next generation of statistical scientist, and a description of how Baylor’s growing data science programs are creating new opportunities for interdisciplinary learning.
Transcript
Derek Smith:
We're talking statistics today on Baylor Connections, but not just numbers, the way those numbers are applied to solve global challenges in water and more with Mandy Hering. Mandy Hering is Professor of Statistical Science, a Baylor graduate, and a leader in this area. And Mandy, we're glad to have you on the program today. Thanks for joining us. We've had you on this before, but now in the video version of the show, it's great to really delve inside the work you do that goes pretty far beyond a computer.
Mandy Hering:
Yeah. Thanks for having me. It's fun to be here again.
Derek Smith:
Well, it's good to visit with this, because you look at the work you do. I think on one hand it seems like it is very much computer-oriented and we all picture in our own math classes, working with calculators or computers. But where your work goes is into municipal waterways and wastewater treatment plants. And I think that only scratches the surface of the work you do. So, let me ask you, where are some of the unique places or unique industries that maybe you never would've expected your work as a mathematician to take you?
Mandy Hering:
Yeah. I think one of the really fun things about statistics is you get to... there's a really common expression you get to play in everybody's backyard. When I was an undergraduate here at Baylor, I was an undecided major for a bit. And I just thought, what am I going to do? Everything is interesting. Every field has a story and everyone's trying to help humanity in some way. And so, how am I going to contribute to that? And so, if you take what you're kind of good at, which was I was a mathematics major here at Baylor. And they didn't have the statistics department when I was an undergraduate here, but they did have statistics classes. Wow, that was really cool in all the different areas you could apply statistics. I didn't realize that before I took those classes. And so, that's what prompted me to get into statistics.
And I never know who's going to walk through my office door and say, "Hey, I have this interesting data set." And so, gosh, data sets from analyzing counter insurgencies, analyzing serial criminal offenses, those were sort of early on. I started out with my dissertation work on forecasting wind speed for wind energy applications. And I continue to work on that for a while. But here at Baylor, there's a lot of people here at Baylor working on water and water-related problems, primarily in the environment. But a fun one is the whale earwax problem with some people in biology and environmental science. So, a lot of things I've even worked on. I can't really tell you what it all means necessarily.
Derek Smith:
Man, that's fascinating. So, you've worked, again, it's like criminology, energy, waterways. For you, how much fun is it to...? You wear your statistical hat, your mathematician hat, but I don't know, like you said, you get to play in other people's sandboxes. How enjoyable is that for you and how much fun is it to be able to share that with your students here at Baylor?
Mandy Hering:
Yeah. I mean, certainly, with students here at Baylor, it's a blend of you have to learn statistics and know it, but the breadth and depth of areas in which you can apply it is really important. And it's funny, because I tell new assistant professors when they are starting out in the statistics field at a new university, I'm like, it's going to be so easy for you to find collaborative partnerships with people. You walk into any university lunch or mixer and you tell them you're a statistician, they immediately want to start talking to you about their data and they've got problems that they want you to solve. And so, I think that's what a lot of researchers really need to begin to advance their own work even further, is someone to really sit down, listen to them, understand where their data comes from, what it means.
And so, when I'm teaching in the classroom, and I have taught undergraduate students, and I primarily now am spending time with graduate students in our PhD program in the Department of Statistical Science, you're learning the statistical methods, but always with an eye to what's the application, what's the goal? What are we trying to do to help people with these models? And at the end of the day, statistics is really just one component of the broader picture. There's been a lot of controversy about p-values and what p-values are in the academic literature. And we're not just chasing "statistical significance" in problems. We really want to see where are there and important differences between groups, one treatment versus another. And so, the statistics and the data that are collected are just one part of that.
Derek Smith:
As we talk, I want to ask you about some definitions, because whether it's p-values or even things that were more familiar.
Mandy Hering:
Yeah, sorry. I went there. I brought it up.
Derek Smith:
Hey, you know what? You are right there in your wheelhouse and we're going to join you there. So, give us a little 101 on what's a p-value, but also how do things like AI, and machine learning, and modeling...? I think these are things that a lot of us hear about. We don't know exactly what they are. What should we know about those?
Mandy Hering:
Well, I would say if you're an everyday person, you probably don't want to know anything about p-values.
Derek Smith:
Fair enough. They can Google it.
Mandy Hering:
Yes, Google it. Well, that's a can of worms, too, I would say sometimes, because even the statisticians in my field that sometimes argue about the pros and cons of p-values, so I won't go there. And my students are going to laugh that we're talking about this, because we spend a few days every semester talking about p-values, and how they're misused, and misunderstood. But I would say across the course of my career, it's been a really interesting time in the world of data-driven modeling. When I first went into statistics, I think the first big buzzword was big data and everyone was talking about big data. All of a sudden, we have the capacity to collect and store huge volumes of data.
And of course, with smartphones, and location tracking, and all these new things that we are able to collect and record. And how do you make sense of all of this information without drawing conclusions or making decisions that are kind of unique? You're missing the forest for the trees sometimes with big data. And so, those were the challenges, the storing it, the accessing it, the analyzing it, drawing conclusions from it. And then after the big data phase, then it was machine learning. We have all these huge data sets. How do we build models for these huge data sets to draw conclusions? And so, there's been a huge proliferation of new types of modeling strategies. Neural networks, and random forests, and support vector machines, all these kind of buzzy, modeling-sounding words.
And then after that came data science. It's like, okay, we can build all these models. What do we do with them? And so, people started talking about data science. Well, we want to be able to use them in ways that we can connect with people in interdisciplinary settings. We want to be able to communicate the results of these models to other people. And then since that, now we've got artificial intelligence. And each one of these components, big data to machine learning, to data science, of course, statistics and computer science play a role in every single one of those, but there's a new twist every time. There's something new that's pushing us forward.
And with artificial intelligence, it's like we have all these large language models now. It's not just numbers, it's words. And we can interact with this large language model, and ask it questions, and get responses that are like you're talking to another human being. And how that impacts how we work and how we learn are questions that people are grappling with now. So, I think it's an exciting time to be... for some people scary, but also I'm a little bit more in the... This is an exciting time to be in this world of data, because we have an opportunity to influence and to guide.
Derek Smith:
It's funny as you describe all that. So much of it, it's high-tech, it's technical, and for a lot of us, maybe only vaguely familiar, and yet you think about the way it's applied. I'm a sports fan. I think I remember when I first realized data could do different things, it was the Moneyball era. I remember the book Moneyball in baseball and how teams were using that to think about building teams anyway. So, that's what I remember. And of course, you apply that across so many different ways, as you said, criminology and wind speed. And what's become a big one for you is water. In fact, I think a lot of us here at Baylor think of you as really working in that water space, first and foremost, even though your career has taken you so many different directions. So, let's focus in on that. When did you first realize that your work as a mathematician, a statistical scientist, could be applied to some of these really big global challenges and opportunities to really serve people? And I mean, when you think about water, that's about as foundational as it gets.
Mandy Hering:
Yeah. And that was one of the reasons that really drew me to water and water applications, is that it's so fundamental to human life, and yet we still struggle with providing it in the right quality, and quantity when and where it's needed. So, to me, it's such an important question that we need people working on. And I really started working on water very shortly after I got my PhD. As I looked back over the course of my career, I'm like, oh, that's actually sprinkled in from an early stage. I minored in environmental studies here at Baylor, and so the Earth and the things that are in it have always been interesting to me, the environment. I started fly-fishing when I was at Baylor, and I just love that. I love being on the water. You think about the very large proportion of people that live within 100 miles of an ocean.
I think it's like 40% of people around the world live within 100 miles of a coastline. Where do we go to recreate? Where do we go to play? Even if it's snow. If you go snow skiing, that's water. But you go to the beach, you go to the lake, you go someplace with a pool, you go to a water park. For me, water just seems to permeate where I want to spend time. And so, I started working with a great collaborator, Tzahi Cath, who's an environmental engineer. And he builds these big systems on the ground to collect water, wastewater, and to study the system, and how it is treating the water or wastewater. And he has sensors everywhere. And he realized quickly, I have all this data and I don't know how to mine it for information.
And one of the very first projects that I worked on with him was really compelling. They had an upset in their system. And we could see the upset coming days in advance with the data, that they could not tell just by operating the system and just visually watching the data on a screen. And so, we were able to synthesize all of that data with a model and give some early warnings, like, "Hey, you should look at this. You should look at this, because if you don't intervene now, you may have a two to three-month recovery period down the road that's going to be unpleasant." And these are academic, educational systems that we're studying, but they have implications for utilities around the US.
Derek Smith:
So, you think about the fact, as we've talked about before, for most of us it's as simple as, "Hey, when I turn my faucet on or the tub, the shower, whatever, I want water to come out." And most often, for most of us who are blessed to have that, where we live here, it happens that way. But what you describe is most of us aren't thinking about what's going on at the wastewater treatment plant or everyone it takes to deliver that water to us that we take for granted. So, take us inside that even a little bit more. You described it here. Obviously, people who deliver our water to us are not immune from challenges. And if a challenge blooms into a real problem, it could really impact a lot of people. So, how important is it to get upstream of that, if you will, which is what you do?
Mandy Hering:
Yeah. I think it's incredibly important. And to be clear to the audience, I am not a water scientist, but I get to work with all of these incredible people who build these systems and study them. And so, I've had the opportunity, for example, to work with the City of Boulder, Colorado. And they get water that we don't think about. It's just flowing into their wastewater treatment facility. And they use oxygen, so they aerate the water to treat, and to take the ammonia, and convert that into nitrogen. So, the nitrification processes is what's important there. And we found models where we can forecast ammonia. Everyone said, "Oh, this is probably not going to work out really well." Well, there's actually a really clear pattern in how ammonia and the concentration of ammonia changed over the course of the day.
And then they could tune how much oxygen they were adding to the water based on those forecasts, in order to, one, treat the water to the proper degree, and two, maybe not over-aerate so they're saving energy. So, we both want high quality water at low cost. So, it's cost. It's energy consumption that also goes into it. And a lot of communities are, even in the US, I think still struggling, especially rural towns. They may not have the infrastructure or the ability to plan in the same way that medium and large-sized cities like Waco and beyond have. And so, I think especially here in Waco, we're very, very fortunate. We have a really great team working at Waco Water to deliver the water and to manage the facilities here.
Derek Smith:
It sounds like they want tools to be proactive. And you're working, you, and your students, and colleagues are giving them more of those tools day-by-day.
Mandy Hering:
Yes. That that's the goal. That's what we hope.
Derek Smith:
And you think about it, too, maybe to state something obvious. I think when we think about some of these challenges, we know the stat. What? 70% of the Earth's surface is covered by water. So, it's not that there's not enough water everywhere, but it's getting potable, usable water where people are. And those challenges continue to grow as cities grow. As you said, 100 miles from a coast is where 40% of the population live. That's a lot of demand that you're trying to help them meet.
Mandy Hering:
Yeah. In a small area. And people think, "Oh, well, desalination is maybe the solution, just to remove the salt from water." That's really energy-intensive. And as we know, the coastlines are places that are in high demand for real estate, so where do you locate those is challenging. So, any sort of big efforts in that direction, it's going to be challenging. And so, I think people are maybe not always cognizant, but there's people working on all ends of the problem, from where do we get the water to what are the new mechanisms and treatment processes to treat it efficiently, to treat it well. How is it distributed to people? Just the infrastructure needed to deliver water to people's homes. I mean, it's just amazing if you think about it.
And then once they use that water, where does it go after that? And it has to be treated, then it's generally released back into streams and rivers. And how does that impact the environment and our interaction with that environment? It's a lot of different people working at different points along that continuum.
Derek Smith:
You got a lot of dominoes to think about as you do this. Mandy, you mentioned you've worked with cities like Boulder and Waco. In your work, you've been supported by the Department of Energy, the NSF. And it seems like a lot of those opportunities have given you a chance to show leadership in continuing to grow the way statistical science education takes place here in looking for new ways to provide opportunities for students. What are some ways you've been able to do that? How have you been able to help train up a new generation to apply stats to these real-world problems?
Mandy Hering:
Yeah. One of the reasons I really got excited about applying statistics on water and wastewater treatment, is it didn't seem like there were a lot of statisticians per se, who were working in this space. So, lots of statisticians usually find a niche area where they're really applying their expertise to biostatistics, genetics, some particular field. There weren't very many statisticians in water. And we realized that because of that, there wasn't as much training, say for environmental engineers or water, wastewater treatment operators, utilities. And utilities were required by the EPA to collect data on their flows and water quality metrics.
And of course, they record data on how they're operating, different operational switches that they toggle on and off. And so, we realized, hey, there's really an opportunity to do two things, to train statistics and computer science students to use data in a really meaningful way, and also to train people who are in these fields to use the data that they have. And so, we were fortunate to receive a National Science Foundation grant. And we titled it Modernizing Water and Wastewater Treatment through Data Science Education and Research. It's a very long name, but we shorten it with... We we call it Mo(Wa)2TER. So, to get more water was the idea of that acronym.
But when we started this five or six years ago, Baylor didn't have a data science undergraduate degree, for example. And so, we started an introduction to data science course, which then got folded into that data science degree that is run by the Department of Computer Science. We also offer that data science course in the Department of Statistical Science and support that degree with several of our own course offerings. But the idea was, let's get students interested in data science. And if we can connect them with a real problem, students with no coding background, no statistics background, but they can come in and they can see, wow, this is a real problem. It's hard. It's hard to wrangle and work with the data just initially.
Even to begin to answer the questions, I have all these other challenges to work with. But once I'm over those hurdles, look at all this cool stuff that we can do, these models we can fit, these dashboards that we can create that people can use. And so, that's the first thing that we did. And then we started running these summer programs, where we'd partner with utilities and consulting firms. And they would give us their data and a problem, and we'd put a small team of students to work together on it. And out of all that, we realized there was a need for people who were already in industry or students who were, for example, in environmental engineering or civil engineering, that they wanted to be able to learn these tools as well, and to go and use them in their fields. So, we have some industry workshops that we developed that we offered both online for working professionals and also in-person for graduate students.
Derek Smith:
Sounds like a real competitive advantage for students to get a jump on this early and work on some real problems, for sure.
Mandy Hering:
Yeah.
Derek Smith:
Well, Mandy, as we wind down, I want to ask... you paint part of that picture already, ask how data science and statistical science at Baylor are growing. You mentioned now there's a major, there's these new programs. As you think ahead to the impact Baylor can have on addressing real world problems through the collaborative research that you and your colleagues do, what most excites you? What do you envision as you look down the road?
Mandy Hering:
Yeah. I think that data and data science has been in small pockets around the university. So, the University Libraries has a data and digital scholarship program that's run by Josh Been. Our department has a statistical consulting services that has been jump started by Rod Sturtevant. And we just hired a new faculty member to come in and support that. And Rod's a great person to talk to. They just celebrated serving their 500th client here at Baylor.
Derek Smith:
Wow.
Mandy Hering:
And they don't just serve Baylor clients, but also industry comes to them. And they've worked with several different businesses as well. And so, I think that what I see potentially, I'm not necessarily the one making all of these decisions, but what I've been seeing and hearing a lot of, especially with Baylor in Deed, is, "Hey, how can we bring all these tools and resources together into a one-stop shop, where students and faculty can go?" So, we're not duplicating efforts in small pockets around the university. Sort of a one-stop shop for all of my data needs. So, I think that's potentially very exciting. Of course, there's a lot of people already thinking about how AI can be used in the classroom. Students are already using it extensively.
And so, sometimes faculty, we're a little bit slow to evolve, but I think that's also an exciting area to think about, how do I incorporate this and use it as a tool? And the College of Arts and Sciences is going to do a data science fellows program next summer for faculty members who... If you go into the consulting center, you can go in with a simple question or a more extensive question. But if you really want to learn the tools that you're going to apply to your data, then you might need a little more attention. And so, that's the idea of these data science fellows, is that faculty members will be selected to come in and really partner with a statistics faculty or maybe hopefully a computer science faculty, and really learn and be able to implement themselves a lot of the tools they want to be able to advance their research.
Derek Smith:
Well, already great impact. And it's exciting to see that continue to grow in the ways that you get to partner with your colleagues. And who knows what industries or disciplines you all will work with next.
Mandy Hering:
Who knows?
Derek Smith:
Well, Mandy, thanks so much for your time today. We're excited to see what's ahead and certainly appreciative of the work that you do in helping deliver clean water to us here in Waco and beyond.
Mandy Hering:
Yeah. Thanks for having me here. It's been great.
Derek Smith:
Great to visit with you. Mandy Hering, our guest today on Baylor Connections, appreciate her being with us and appreciate you being with us today. A reminder, you can hear this and other programs online, get video, audio, and more at baylor.edu/connections. I'm Derek Smith. Thanks so much for joining us on Baylor Connections.
