Decoding Data Literacy: Perspectives from Colonel Nicholas Clark

COL Clark - WP

Dean: [: 00:00:00

Colonel Clark is a 2002 graduate of West Point. He's currently serving as an associate professor in the Department of Mathematical Sciences, where he teaches a number of courses, including statistics, mathematical computation, and calculus two.

Before returning to his alma mater as a professor, Nick earned a PhD in statistics from Iowa State University.

Nick, welcome to the podcast.

Colonel Clark: Hey, thanks sir. Appreciate it.

Dean: So let's tell, tell me a little bit about your background.

Colonel Clark: Sure. So I, ironically enough, sir, when I first entered West Point, I thought I was gonna be a law major.

Dean: You threw your life away, went all into math.

Colonel Clark: That, that's right. But I ended up being a, a math major and I. I, when I was a cadet, I was doing as much theory as possible. I wanted to do kind of the most abstract, my first year research was in music and mathematics and how they blend together.

bout to go over to Iraq. And [: 00:01:00

And it just dawned on me, we were making decisions based off of someone's feelings rather than the data that was sitting in front of us. So using what I had learned in my MA 3 76 course, I just built out some simple regression models. I went and presented that at a symposium and they said, No one's ever done that before.

And it, I've really started to see in my career that it was only the statistics I was using. I wasn't using any of the theoretic math. So when I had that choice to come back and teach here as a junior faculty member, they said, you can get a degree in anything you want as long as it's operations, research, or math.

And I said, well, what about statistics? I said, oh yeah, maybe that too. So I ended up getting a master's in statistics. And the rest is just, so we're gonna

is practical application for [: 00:02:00

How does that, what's your trajectory been?

Colonel Clark: So after I went here as a junior faculty member. I left here, I was still a military intelligence officer. I went back to a special operations unit.

They like to say that I was the only card carrying statistician. So I sort of had a dual role of, of data analysts and statistician , for that organization. And I really start to see, and this goes to some of the data literacy work that we doing, I really started to see that over the years, and I think a lot of this was born out of the data that was collected in Iraq and Afghanistan is our senior leaders were starting to say, Hey, we need people that can analyze this better.

So my job became less and less of the traditional military intelligence, more and more of the, Hey, I've got this stuff that I've been collecting for a while. Can I do anything with it?

Dean: How, how did you become an associate professor at West point?

ck here as a senior rotating [: 00:03:00

When I, after I had left as junior faculty, I had a great time here, and I made that tough decision that a lot of our faculty members make as to whether I wanted to go the command route or whether I wanted to, to come back and serve as a senior faculty member here. So I went and got a PhD in statistic at Iowa State.

Went back to West Point, was the director for our Center for Data Analysis and Statistics was selected to be an academy professor. And recently last year I had enough stuff that I had done, put together the packer for Associate professor, but that really was a mix of not only the academic work, but also the research and some of the other supporting Army operations that I've done since I've been a, a faculty member here.

Dean: Okay, so, tell me a little bit about your center?

lytics has grown. It's, it's [: 00:04:00

So right now the center is doing a lot of work where they would go out and find projects that the Army is working on where our faculty and our cadets can potentially help on, they bring them back, work on year long projects AID's going into their thesis year long thesis projects.

Dean: let me just ask you a few very, very basic questions.

When we say data, what are we talking about?

Colonel Clark: It seems like it's a, a simple question. It should have a, a simple answer, straightforward answer, but really, data can take on a whole bunch of different forms. Traditionally, what we think about data is something that, that. It's in a spreadsheet, right?

You have zeros and ones. [: 00:05:00

You know, for instance, we take sound and in its raw form, you might think, well how is this data? Well, okay, maybe you got the frequencies, but can I also take the the sound and visualize it? And then can I take how a computer would look at an image and turn that image into new knowledge? So data can be really, really broad and wide, and that's one of the things we really want to get from our crets here, is not only comfort working with traditional data sources.

different than anything you [: 00:06:00

Dean: And so what do you mean when you say someone needs to, young officers need be data literate. What do, what do we mean by data literacy?

Colonel Clark: It's a great question and as, as you know, the idea of data literacy training has become huge in the army. Really proud of the work that we're doing at West Point in, in helping educate the force with that.

I like to kind of just paint an equivalence here if, if I think about my data workforce, like my medical workforce. What I'm doing when I'm building out my applied statistics and data science majors is I'm sort of building out those, those platoon medic. And I need one of those for every 40 soldiers.

ger before I went out on the [: 00:07:00

Force data literacy is the idea that every single officer, every single soldier before you get out and may potentially interact with data, let's give you those best practices so you can help enable that larger work data workforce to do their jobs.

Dean: Yeah. So tell me a little bit about the data literacy 1 0 1 course that you developed and why you developed it and, and, and how you see it helping the army.

Colonel Clark: As you're aware, we at West Point have operational experiences, which is sort of our idea of sabbaticals for our academy professors, where we can get back out to the force and, and provide some intellectual capital as well as see what's going on out there and see what types of problems that are occurring out on at the edge can be brought back to West Point.

ur special operations units, [: 00:08:00

And I started doing that and until the Afghanistan withdrawal happened. When the Afghanistan withdrawal happened, what we found was the data that was being provided was really, really messy. And I had a, a number of folks who were working with me who are really talented data scientists, and they never actually got to the point where they were doing data analytics with it 'cause they're spending all their time cleaning data.

And we didn't really even understand what questions they want.

Dean: What do you mean by cleaning data?

Colonel Clark: Sure. So, you know, we were talking about in the beginning, the data can take on a lot of different forms. You know, sound can be considered data.

I can't just take though the words I'm being said throw them into a regression model, I gotta change them in some ways. So it's, it's that necessary pre-processing or working with data to get data into good format so I can apply algorithms against them.

Dean: And there was [: 00:09:00

Colonel Clark: Yes, sir. Yeah, so I, I just asked the question, what training are we providing to the force. And there was sort of this idea, well, we got these Coursera courses, we got these data camp courses. And I said, okay, what's the completion rate of those?

You kind of get this head down look where they don't really wanna tell you that nobody was completing those courses because it wasn't, it wasn't relevant to the people that were taking 'em. And we were starting saying, do this on your own time too. So I asked, can I, can I develop a course that gets at the fundamentals of reading data, working with data, analyzing it, and communicating effectively from it?

making people data analysts, [: 00:10:00

You need to ask good questions. You need to properly format data. You need to ensure, you know, the strengths and weaknesses of some of the algorithms that people are gonna use for you. Wow. And most importantly, when we get the results, we need to know how to communicate this in the most effective way so we can actually create a change in our organization.

And so that's what we did. And, and it really took off outside of even the, the organizations I was working with. Somebody at at the Army Talent Management Task Force found out we were doing that and said, Hey, would you mind if we send this out to the Army? I said, sure. And that became my job for the remainder of the OE is just traveling around given the training to, organizations in the Army.

to help us export this, this [: 00:11:00

Colonel Clark: So one of the things I think we do well at West Point and, and admittedly I'm biased, right?

But I think we educate well and one of the reasons we educate well is we talk about pedagogy or we talk about how we teach and we've got people that really are good instructors. And this course really works the best. Even the, even if I kept the course how it was, and, and Nick kept traveling around given this training, it's gonna get stale after a while.

h them how to put together a [: 00:12:00

'Cause now you can go and you can give this and give this on demand. But you can also have some staying power. You can look and start to analyze whether the, the work that I'm doing in that organization has actually moved the needle a little bit. Have I improved my overall data literacy for the organization?

So the next thing we're gonna give them over this 40 hour course is a framework for them to assess their unit's level of data literacy. So now you're able to deliver this training and you're able to both check yourself as well as perhaps we can build some relationships, we can travel around and, and we can evaluate each other.

the course being implemented [: 00:13:00

Colonel Clark: I think first is ensuring that this doesn't become just another training that you have to do. It can't be something that now is the data time, everything else is the non-data time. And we go and we get this and everybody falls asleep in the back of the class. We get your annual data literacy refresher.

That's not gonna work for this. So I think the challenge is once we start to provide this, actually implementing it, and, and this to me is a command responsibility where if I've had my organization that goes through this and they start to brief me on something, and I know because I've, I've got some level of data literacy.

his process works. Let's get [: 00:14:00

And when those commanders start to enforce analytics on their organization, enforce when they come to me, it's the old Deming quote in God, we trust L LPR data. Bring me the proof behind this.

Dean: Now you you, you're hitting on an interesting point. You're saying that basically what's gonna drive this though is command emphasis.

Yes, sir. Have you found that senior officers are, are resistant to, to, to the application of data or being, are they flexible enough to start to move towards also becoming data-centric and how they operate?

Colonel Clark: I think at, at the seniors level, we're starting to see a lot of remote, I should say. We are continuing to see a lot of encouraging signs.

ry Command organization. I'm [: 00:15:00

Where sometimes they hear the guidance and they don't really internalize quite yet how to implement it. And I think the more that we flood the system with people who are getting this training, and as we were talking about before, seeing the trade offs directive that data literacy starts to, to be incorporated into the professional military education will only further to enhance this.

So I think it's moving in that direction. I think with, with anything, when we're talking culture change, 'cause that's really what this is. It's not training, it's, it's a change in culture. It's gonna be a little bit slow and you're gonna see sort of fits and starts. But as long as we keep focusing towards that direction, I think we'll get there as an organization.

Dean: So what do you see as the future of data in the Army and the Department of Defense?

Colonel Clark: I think it's just gonna grow, sir.

t to see more of a demand for[: 00:16:00

We always do need to keep an eye on exactly what you brought up before, though it can't, we can't let this pendulum swing so far that we completely discount any sort of experience, any sort of subject matter expertise, 'cause that's what really makes this work. It's not just a bunch of us nerds who are sitting coding.

It's a combination of having people that can speak the language of data with subject matter experts who can really contextualize and, and make this thing work

Dean: So why is, and you've, you said it a little bit, but let me just ask you more pointedly.

hy is data literacy, in your [: 00:17:00

Colonel Clark: I think, you know, more than anything else, where we sit right now as, as an army, is that every single officer has that ability to interact with raw data. As we just said, data is, is very broad as to what this means. But when you're out in the middle of, of nowhere and all of a sudden you have a sensor that was.

You know, given to you from somebody from a high, higher headquarters. Well, now whether you know this or not, you are part of the army, state of workforce. You are interacting with that data, and you have to figure out what you can do at your level to triage this. It's not enough to just kind of take this and, and throw it over the fence and wait for somebody at a higher headquarters to, to analyze it for you to send it back to the subject matter experts, contracts that you've hired.

s, that we've structured the [: 00:18:00

Dean: Would you agree that data isn't a, a panacea, it can't solve all problems, right? There are situations, particularly in combat, where the fog of war creeps in and, and we expect our officers and, and NCOs and soldiers to be able to adapt and adjust very quickly. And so what is the proper balance between relying on data and, and still relying on the art of leadership or, and the other way, and the other way to look at it's how do we ensure that commanders are not deferring responsibility for decisions by being over-reliant on data?

is that it's not necessarily [: 00:19:00

Everything needs to be data informed. So I, I like it better when I say it's data informed decision making rather than data-driven decision making and kind of coupled with that is the leader responsibility to understand the limitations and the strengths of the data that you have sent in front of you.

It's not enough to just rely on, on your, your operations research analyst who's sitting at your headquarters, who's got the training at this, at that leader level. If you don't understand, you know, again, use the, use the example that's everywhere right now Chat GPT. If you don't understand the strengths and limitations of this algorithm and you just blindly use this, you are not truly part of that data literate workforce.

Dean: So our listeners. I won't know this. Sure. And it's a congratulations, but it's also an indication of, of how far this has come.

nd dollars from the National [: 00:20:00

Colonel Clark: No, this is, this is, this is a lot of fun. Uh, Was some joint work that we've been doing with Carnegie Mellon University, Pittsburgh, Baylor, Yale, and St. Lawrence University on how do we engage an undergraduate population in data science better than what we currently are doing. And specifically we wanted to look at whether using sports statistics can help teach some of these lessons that perhaps, you know, I was just in the classroom and I was using engineering loads from a given week for some imaginary engineering system.

For the case study, for the, the, the algorithm that we were using. Like, wouldn't it be so much more powerful if instead we were teaching based off of something the student actually cared about some data that they actually were interested in. And I actually got into this because one of the faculty members at Yale used to be a faculty member here at West Point.

great things about our, our [: 00:21:00

And so we kind of got to do a little bit of that and he said, let's make this bigger. Let's look at all sports. So we're building out these modules right now where if you wanted to learn something about linear regression, you can go onto the website. You find a sports module. We have these people know a lot of people for, so we get some professional athletes and people in sports industry who will open with a video as to why this concept's important for them.

We teach them a little bit of stuff, give 'em some, some environments that you can play around with the data, where you don't have to code a lot, but you're getting those key skills that you need to learn through the lesson. And then you tie this back at the end, you get the video comes back again.

ork as analysts for an N H L [: 00:22:00

Now we're gonna build a process in place where if you wanna build one out and you say, Hey, I really like, Broom Ball. I know you're a big Broom Ball guys, sir. And, and, and I want to teach something about linear aggression. You contact us, we put you in, in contact with a sports professional who, who works in that discipline.

We help you build this out, and then we publish it. So it's widely available for everyone.

Dean: And you've found cadets are fired up about this?

Colonel Clark: I, I, again, I think just like getting Army applications in front of them, sports is a natural thing that cadets are just gonna get fired up about.

Dean: I can't help but think of about Billy Bean and Moneyball when you're talking.

nections to using data to be [: 00:23:00

Do you find there's an increased demand from cadets to be involved in this, this field?

Colonel Clark: Oh, absolutely, sir. In fact, the class of 26 just has our highest number of applied statistics and data science majors. We have 20 people that selected that as a major. And even the folks who aren't part of that major, more and more of them are doing research that focus really around data or algorithm development or kind of computational math.

Dean: Why do you believe cadets are demonstrating such you know, interest in this particular field? Whether it's, data or applied statistics or, or what else?

Colonel Clark: I think right now, if you pick up a newspaper, for instance, Chat GPTs everywhere, so we get examples of these algorithms that are out there and, and a lot of our cadets want to go that next level beyond it.

ce major was really designed [: 00:24:00

Dean: So you found almost by accident the use of your math background and data and your time in the Army, it's led you into, into becoming one of the premier faculty members in at West Point that is in this intersection between math and data. How has that driven what you have done here at West Point and how you're trying to prepare the next generation of, of young officers to be, but not just data literate, literate, but data centric?

cy in their research as well [: 00:25:00

There are cadets and faculty right now that are doing actual data science work that is directly supporting our operations forward. And I think by having that, that experience, by being able to demonstrate to the cadets that, Hey, this isn't, this isn't something that's sort of esoteric as much as I like to do central limit theum proofs.

That's not what this is. That's right. It's, it's taking data and making a good decision from that data that you've got collected in front of you.

Dean: You, were instrumental in developing West Point Supply Statistics and Data Science major. Why was this a priority for you? When you, when Yes, sir.

And whether that was when you were a rotator, you recognized it, or when you came back as a, as an academy professor. Where, where did this become a priority and why was it a priority?

Colonel Clark: I think it was really around the time I was working on my PhD and it was in the growth of data science when I started to see that.

What people [: 00:26:00

But really over the last decade, and I saw this at Iowa State, people were doing more computation. They were doing more non-traditional things. They were taking really interesting problems and trying to solve them. And a lot of this was due to the growth of just data in general. So as the growth of data grew up, what we were able to do with it changed and, and how we had to go about and analyzing it changed.

sts in the army too. Mm-hmm. [: 00:27:00

We're not actually gonna do anything with it. So we kinda looked across the, across the discipline and said really two things. One is there was a need there. There was clear shortage of folks who could really communicate effectively from data, who knew how to, how to answer interesting data questions.

Knew not only how to build on an algorithm, but why those algorithms work. But there's also this change in academia itself, and whereas data science kind of was birthed out of industry, it has taken on a new life form inside of academia. As you are aware, right now, we're seeking a better accreditation for our, our applied statistics and data science program because there are now academic standards for the discipline.

eed. So we should do it, but [: 00:28:00

Dean: So finally, for anyone that isn't mathematically inclined, but wants to be more data literate and isn't able to attend your course, What, what, what could someone do to, to self-develop?

Colonel Clark: Yeah. Great. Great question, sir. I think until we get these modules up, you know, once those are up, I think those will serve as a, as a great starting point. If you're in the military, encourage you to talk to people in your organization. We're going to the training outside of the military. There's some, some books I like.

e more you can Read gets you [: 00:29:00

Download our python. It's, it's scary the first couple times you do it, you're gonna fail. But you know what? Computation is cheap. It costs nothing. Just try something out and, and the pain points that you feel will be your unique pain points and things that aren't working for you. And you, you won't know that though until you've started collecting some data and start actually trying to, to work through some project.

Dean: Well, thanks. Thanks, Nick. Lemme go off subject a little bit. Sure. And lemme just ask you a few other questions. Okay. So what do you, what do you like best about teaching at West Point?

. But in that classroom when [: 00:30:00

Dean: What's the biggest difference you'd say between cadets when you were a cadet and cadets now?

as we all make it out to be [: 00:31:00

They can do analytics better. They can do theory a little bit worse. Like there, there's just sort of this trade off that's gonna happen as, as society changes. But the, the core person that signs up and says, yeah, I'm gonna leave at the age of 18 and go serve my country, is, it's the same. It's the same.

It's, it's inspiring.

Dean: Yes, sir. Hey how many digits can you recite in PI

Colonel Clark: 24? Prove it. 3.1 4 1 5 9 2 6 5 3. Five. 8 9 7 9. 3 2 3. 8 4 6 2. Six four.

Dean: It feels like more than 20. 24 I think. Oh, it was 24. What? Found

Colonel Clark: four And my teacher used to have these up in my, my junior year calculus class. And I would just get bored and I would just sit and memorize it.

to tune into the inside West [: 00:32:00

Remember, you can find this podcast as well as the other podcast journals and books hosted or published by the West Point press@westpointpress.com. Until next time.