Data science and artificial intelligence have inescapable influence and power in our world. The people who are the most negatively affected are often the ones whose voices are not heard. What does a digital world that works for everyone look like? And who gets a seat at the table?
In this episode, our hosts Colleen Ammerman and David Homa speak with Dr. Brandeis Marshall about the consequences of data inequity, the balancing act of qualitative and quantitative mindsets, and the critical importance of humanizing data systems. Brandeis is the CEO of DataedX, a data science consultancy, among many other pursuits. She develops tools for practitioners to interpret the racial, gender, and socioeconomic impacts of data and technology.
Read the transcript, which is lightly edited for clarity.
Colleen Ammerman (Gender Initiative director): So, today, we’re joined by Dr. Brandeis Marshall, CEO of DataedX, a data science consultancy, among many other pursuits. Dr. Marshall develops tools for practitioners to interpret the racial, gender, and socioeconomic impacts of data and technology. Welcome, Brandeis. Thank you so much for joining us for this conversation.
Brandeis Marshall (CEO of DataedX): Well, thanks for having me. I’m really looking forward to this.
David Homa (Digital Initiative director): Brandeis, good to see you again.
BM: Nice to see you, too, David.
DH: We’re going to go right to data equity — big umbrella — and talk wide. So, Brandeis, one of the things we want to start with is what exactly is ‘data equity’ and what are some of the consequences of data inequity?
BM: Right, so, let me start with what ‘data equity’ is. Data equity means that your data is inclusive of people of all backgrounds. So, that is race, gender, and class. That means that the data itself is doing its best in order to guard against any biases. And, this is very difficult because every time you are trying to grab data, you are putting your own past experiences into it, right? There could be some features that you don’t even know about. There are certain things that you have a skewed view of and this affects how you collect data. So, ‘data equity’ is really trying to take a step back, thinking less quantitative and trying to think more qualitative. What are we trying to attain? How is this data actually trying to move things forward? So, data equity is really trying to move forward quantitative people in a way that’s a little bit more qualitative, which means let’s start thinking about the people on the fringe. Let’s start prioritizing those people that are most vulnerable. Let’s now try to figure out what is our impact first and not just try to do things because we can do them.
The way data inequity has manifested in our society is, well… look anywhere. There are spaces where… okay, so, I just voted. I winded up in a county that’s majority people of color, majority Black, to be very specific. I waited in line for six hours and twenty minutes. But, in a county that is majority white — so all the “non-people of color” — the wait in the line was about fifteen minutes. So, to me, that’s the idea of voter suppression. That is the point of, okay, all of a sudden there is no idea that there’s a lot of people coming to the polls? There’s not enough polling stations, there’s not enough poll workers in [these] places. There’s disregard [for] the fact that people are trying to do their civic duty. To me, this is all going back to data inequity. If you know the population of a county, you know how many people are registered to vote. You know that there has been a big campaign in order for individuals to be voting. You can expect that there’s going to be a lot of people wanting to vote, especially if it’s in-person, especially if it’s the first day. And then, as far as the reporting has come out… people started to get in line to vote at about 4:30, 5:00 am. And the polling station didn’t open until 7:00. There are plenty of opportunities to know [with] just small data points and how they have connected to each other, to know that this was going to be a long line. However, there was no type of forecasting, there was no prediction. This is when data could have been used more equitably in order to then be able to deal with the influx of people. So, that’s just to give a very small example. But this happens in every sector of our society and it happens on a minute by minute, hour by hour, day by day basis. And, for Black people, in particular, this is the way we live. It’s just not fair. [I’m] not saying everything in life should be fair, but this is literally trying to push down, push aside, keep marginalized certain groups of people. And, it’s very intentional and it’s very strategic. And so, data inequity happens a lot. What I’m trying to do is push forward the agenda of having more data equity. To not make those people on the fringe feel like they’re on the fringe and actually include them as part of the main thrust of our society — because we all are equal, but we all don’t have equality.
DH: What is it about people who work a lot with data that makes them not realize that they should look at this more qualitative side of it?
BM: Okay, I’ve been in the computing field for twenty years. I’ve taught for the majority of that time, since I was a graduate student. And, what I can say is that when certain people are analytically minded, they have what I would [say is] akin to a fixed mindset. And they believe that numbers are facts and that numbers are the only thing that can be trusted. And so, these individuals truly adhere onto the idea that any time that you are creating something inside of a digital space, the intention supersedes the impact. So, they believe that the intention can somehow be kind of thrust into everyone’s understanding, and that is what is going to help everybody. And, it’s not — it’s not the reality, right? It’s just not the way it is. So, it’s very hard to shift mindsets when they’re very determined that numbers are the solution, that facts are going to get you to a plausible end.
What I try to do is shift a little bit in how I, you know, speak to them and how I give them examples, because I think it’s important that they see, “here is some bad impact.” So, case studies — I tend to use. With those individuals they will say, “Well, that’s only one case.” So, then I bring up another case. And, [they’re like], “Oh, that’s just another case.” And, I say, “How many cases would it take?” How many times have you been writing code and you get an error? You don’t understand what the error is. You call it a bug, but you find out it really was a logic issue. So, how many times do you go trying to fix an error? And, then they will say, they’ll keep doing it, they’ll keep doing it ‘til they get it done. Okay, so why don’t you take that same approach when it comes to the impact of your work. And, that tends to really set them off-kilter. It’s an uncomfortable place for them.
So, then I say, okay, now that you are kind of getting that what you do in code is what you should do in the real world, then it’s just a matter of continuing to have a conversation. And, then it kind of spins out, because all of a sudden it’s like you have a little child who just realizes [that] they understand their ABC’s. They just want to suck up everything. But there is definitely a difficulty in having people who are very analytical understand that there is so much value in qualitative approaches. That, analytically, you want to think about what’s happening for the masse, [but] the people, and the items, and the ideas that are on the margins actually are the ones you need to think more closely about. And, that’s where I think qualitative approaches are so much more important, especially in this time when technology and people and data are all in the same soup of the world.
CA: That is so interesting. I mean, it sort of sounds to me like what you described is trying to dislodge a sort of ideology or a sort of belief system, right? About, like you said, the veracity of numbers and what information you should rely on. Like you said, it’s this fixed mindset and you’re trying to loosen it up and get people to think in a new way, which is really powerful.
BM: Yes. And, it’s very, very difficult. Because, you understand, ever since you were five years old, four years old, you were inundated with math, reading, and they’re separate. And, you either are a reader or you’re either a math person. Like, you can’t really be both. Well, I was both. I danced when I was younger. So, I took ballet class, and jazz, and African [dance], and clogging — I did all that stuff. And, I also loved math.
So, for me, I see the duality of the creative brain and analytical brain and how they need to be synergistic. Because otherwise, we’re not a whole person. And, it’s the same thing that happens inside of tech, inside of people who are coding, inside of how you manage data. You can’t be one without the other.
CA: Yeah, that makes a lot of sense. In this duo, I’m the qualitative person and David’s the quantitative person. In case you were wondering. In case that wasn’t clear. [laughter]
BM: I figured that out. [laughter]
CA: But that is in fact one of the things that we really enjoy about working together. We come from these two places, but then we have a [shared] interest. And, like you’re saying, [we’re] thinking about it more synergistically.
It reminds me of a question that I had — maybe a different way of tackling the same question you were just talking about. I was thinking about taking it up a level. It seems to me that obviously technology, which is such a huge catch-all, is top of mind for individuals, companies, institutions, and governments. And, I’m always kind of struck by how there are a lot of conversations about inequity in technology, things like algorithmic bias and how homogenous the tech sector is, but then there are all these; other conversations that are technology-rooted, that are also very vital — like cybersecurity, right? Like the power of big data. These are very top of mind for companies, for boards, for things like that. And it seems to me that [the conversations] are often very separate. So, even the issues around technology inequity, they still kind of get siloed. But it seems to me there’s got to be a relationship between topics like, say, cybersecurity and [equity] concerns. So, I’m just wondering, going up from the individual to the macro level, how you see that. Why do [these conversations] get separated and what can we do to have a more integrated conversation?
BM: Yeah, I think that there is a lot of grassroots effort, right? There’s different organizations. There’s non-profits. There’s, of course, for-profit companies that are trying to make these connections between the two. And, I think, quite honestly, capitalism is making money the top priority. So, when money is the top priority, people aren’t the priority. And so, making things siloed is actually really good for business, right? Because you have one track that’s working on one problem, another track with another problem, and finding connections between the two doesn’t necessarily optimize the bottom line. [What] would optimize the bottom line is to create as many threads as you can in order to get money. But in order to solve the problem, they want to now create avenues in order to solve it a certain way. Because that sort of way is not going to detract or mitigate or disrupt their funding channels. The trouble is, in order to really make impactful change, to make an inclusive work environment, to make an inclusive product, you’re going to have to work across different sectors. And so, that means that companies and organizations are going to have to now cross-pollinate. There’s no other way out of that particular box. If you really want technology to be inclusive, that means you have to do more than have a diverse team. You’re going to have to actually make sure there is diversity in leadership, which means there’s a power structure that’s going to need to change, right? That power structure means you’re going to have to make people who are, quite frankly, white supremacists or act[ing] for white supremacy at the very least, step back and stand down. How is that going to happen when the money is the ultimate goal?
DH: So, I wanted to talk a little bit about raw data. And, data historically can be very messy, especially as we get older. And, a lot of people don’t really understand what that means when you try and infer or learn things from that. Is there a good way that you would recommend people think about the underlying data underneath systems that are trying to advise them?
BM: Oh, that is a good question. Data is everywhere. So, how we bring it into our own orbit is what is always interesting and that can be problematic. So, I guess the best way to describe it is to really look at it in the eyes of a child. When a child learns something new, what happens? There’s all of these videos about the first time a child, like, eats ice cream and it is hilarious, or the first time a child, you know, understands what a popsicle is. That is what happens with raw data for us when we receive that raw data. We don’t necessarily know how to navigate it. We’re a little stunned by it. So, every time we’re in a system, that data has been in some way sanitized. And so, we have to kind of think back and recall — and I know it’s hard to do — what it was like the very first time we actually saw that content, we saw that data. And, data comes in many different forms, right? It’s audio, it’s visual, it’s a culmination of many different things. But the data, the rawness of data, is very helter-skelter.
CA: It’s interesting, what I heard you saying, Brandeis, is that we need to bring a certain amount of humility to encountering data. I think there’s an interesting parallel to what you were saying before about institutions. We need to bring some humility to what we encounter and not sort of think that we have the answer a priori, or even have the framework for understanding it, right?
My only real work with data is with this longitudinal study of HBS alumni that I’m part of. And, I’ve learned a lot there. I’m not trained, aside from taking Statistics 101 in college twenty years ago. One thing that I learned from the people on the team who actually can run regressions and who understand data is, you know, oftentimes I’ll make an assumption looking at a table or something, and it might be interesting. But it’s easy for them to point out, well, here are some things that you are not considering that sort of change the course. As somebody who likes to impose a narrative on things, as somebody who is a reader, it’s been helpful for me to realize that. What you’re saying resonates because it is about having that humility and knowing that even if you have a story in your head that’s very compelling, you can’t necessarily impose that on the data. There may be things that you need to cultivate the ability to see and to be aware of, even if they kind of disrupt what you would like the story to be.
BM: Yeah, exactly. Exactly. And, that’s the reason why having groups of people look at the same data with different perspective, different lenses, is so vital. That’s the reason why having people who are on the analytical side and also people who are — I love your words, you said, “I’m a reader.” I have people who are more on the communication side actually look at the same data because we see it differently. And, that’s very important to get to the same conclusion: to see it differently, different paths. And, that’s okay. That’s normal. That’s human.
DH: That makes sense.
CA: Just to be clear, David does read — he sends me articles constantly!
BM: Well, that means that he knows how to send things. [Laughter]
DH: That’s true. I can’t send things to Colleen without having read them because she asks me hard questions about them. [Laughter]
It’s interesting. I’ve written a little bit about the future of workers, this ideal of someone who has a balance of computing skills, applied math skills, and domain expertise. I wonder if you could talk a little bit about what are good ways to bridge domain expertise, because when it comes to data, you have to sort of know something about what you’re looking at. If it’s, for instance, weather data and you don’t understand how temperatures work in different parts of the world, it’s very easy to misunderstand it. The same is true of data about people. For instance, I know that this is a big question in social media where online use of language, what is considered inappropriate in some places and in other groups is not inappropriate. The use of words is different, how they’re used. How do you think about the future of bridging that divide? Is it just people literally with different perspectives or is there a way to get that into one person?
BM: So, I’m of the philosophy that it’s not possible to get it into one person. I think you’re going to be highly leaning toward one of those three pillars, right? You’re going to be in that domain. So, I think it’s very, very important that you keep your mind and your work open to other people’s perspectives. Because that is where there’s strength; there’s strength in the different thoughts. There’s also strength in, not necessarily compromising, even though you are compromising your own ideals — you are trying to grow yourself. You’re in this growth mindset. And the only way to do that is to be around people who are not like you. And, the more that you spend time around people that are not like you, the more that you realize the different ways in which data can be easily manipulated. And so, you wind up being more human. You’re developing your humanity, in a way. And, that is to me, is what tech, as the big umbrella, is missing, is the humanity of it all.
I think tech is just starting to get there — this conversation around the data, and the fact that data is a commodity at this point. It’s being traded like, I don’t know, casino chips. I think it’s important that more people have an awareness of how their data is being used and how not one person can be all things. And, I think that happens to be the issue in a lot of organizations. They want a data scientist who has twenty years of experience in like five different disciplines. That’s not going to happen. You’re going to need five different people, dang it! You’re going to need five different people. Like, this is not going to work. But again, going back to my earlier comments about that ROI, that money at the end — if you can hire one person to do five things, why hire five people to actually get you to that goal you’re looking for? So, there needs to be that type of shift. No, it needs to be multiple people. It needs to be multiple people.
DH: And, those multiple people, you said in there — they obviously have to have an appreciation of the other sectors. And, I think that’s where we see sort of this siloed world where people aren’t trained necessarily, or even have experiences, like you’re saying, workspaces, to gain an appreciation for those other spaces.
BM: Exactly, exactly. And I think that’s part of the issue, because work has been very much, “You do your job, you stay in your lane,” that means that the silos just kind of reinforce themselves. It’s hard to get out of that grind. That’s the reason why I think, at least for now, while we’re in this pandemic, which means people have to be online, which means they’re trying to find connections in different ways, this is an opportunity for organizations to now try to do ‘cross-pollination-ships,’ to really start to think through: “How are this team and that team working toward the same common goal? What is really the strategic plan? How are these different parts actually moving together and are they moving together in the first place? Are they designed to move together? How do you design them to move together?”
That’s part of the work that DataedX is trying to do — to be that strategic partner in what is happening in your data practices. Where are there silos? What type of skill toolkit do you need in order to bridge this gap? Do you want to really bridge the gap or do you want it to sit on a shelf? Do you want to be inclusive? And, how do you now think through some of your operations in a way that is looking at data to be informative, to be data-informed, not data-driven? To really be data informed, to really understand what the data is telling you in order to then make actionable change within the organization. And how can we strategically get you there? Is it talking to the data analytics team about algorithmic bias and fairness? Is it then going to the front-line workers like the cashiers and just talking to them about, okay, how do you engage with customers and get information from them to make sure that their address and their email is correct and let them know of different reward systems? There is a lot of richness on how we move toward the same goal. But we do need a more holistic perspective.
CA: I just love everything you’re saying about the growth mindset piece. I hadn’t thought about that connection, but I totally agree. It’s easy for somebody like me to say, “Well, I’m just not really great at math;” to have a fixed mindset about it. Instead of saying, “I have the ability to actually understand more about what’s out there than I believe.” Anyway, I just think that’s a great connection.
BM: Yeah, it’s really the way we’re taught things is also an issue.
CA: You’re talking about how people are educated, who is educated, these systems that sort of determine who does what, basically. And so, I just would love to hear you talk a little bit about what are some of the ways that we can create that change. It’s very easy to say, yes, we need more underrepresented groups in tech. But what actually, as you see it, are some of those critical pathways for getting folks who are not dominant in that sector and in that field into the industry in a more effective way?
BM: Yes. So, I think a lot about this. And so, I would say, number one, that internet connectivity and machinery equipment needs to be priority. I think that at this point, how connected our world is, how much we rely on systems that are digital — that needs to be in every nook and cranny. So, there are places in this country that do not have good connectivity to the internet and that needs to change pronto, because if you do not change that underlying infrastructure, that everything else is for naught, right? There’s no need to be giving computers to kids if they have no way to connect to the internet. And so, that’s number one.
Number two, it would happen to be a lot of education on how to use technology. And, that means what are some of the pitfalls in doing it too much? What are some of the pitfalls of not doing it enough? When are there good times to use technology? Like banking is pretty much online now. So, having that type of instruction of fiscal responsibility and financial literacy is important. And, I suppose start with the parents and guardians of children.
Now moving forward to actual adults — that means that now it’s going to be about when to replace technology. That is something that is not really discussed at all. When do you move from one mobile device to another? When do you know you’ve used too much of the memory? How do you know what is on that particular device when transferring it? So, those are just some of the infrastructures.
So, in talking directly about technology, the same thing needs to happen inside of tech. Something that I do is amplifying Black people inside of technology — those that are in the past, those that are currently there today And, we do this amplification a lot. In the past few weeks, there’s been — past few months, excuse me — there’s been a lot of “Black in” weeks. There is “Black in Computing,” there’s “Black in Engineering,” there’s going to be “Black in Data,” “Black in Cancer,” “Black in Cybersecurity,” [etc]. This is one effort inside of the Black community in order to amplify ourselves. But why do we need to amplify ourselves when we’ve been here all along? There’s a number of different ways in which we can now build that function, so, that being a part of this industry means that you’re actually included. Because, the way that things are going right now, the way things have gone, is that if you happen to be a Black, Hispanic, Indigenous person, you don’t really belong in tech. You can get the degree, but you’re pushed out within five years. And, that is what’s in the literature. And, that is intentional. That is strategic.
DH: I wonder, what set you on your computer science career?
BM: Interestingly enough, I talked about this a little bit before, it was a Gateway 2000. Who remembers the Gateway 2000?
DH: A cow box.
BM: Yes. Yes. [That] sets everybody apart. [laughter] You’re like, what generation are you part of? That was, for those of you who don’t know, the very cheap machine back in the 90s. My father and mother decided to get a machine, a desktop in the house, put it in the very cold basement. I was interested in the icons. And so, that kind of spurred my interest. I knew I wasn’t going to be a dancer because I am vertically challenged. I’m a people McNugget. [laughter] I really, I really was interested in what that [computer] was about. And so, that really set my course, and I was really good at math and I enjoyed numbers. And so, I said, well, what is this coding thing? What is this computer science? This looks kind of cool. And so I just pursued that in college. And then as I finished college, it was the dot-com boom, and something in my gut was like, “This is not going to last forever. This is not going to last forever.” And lo and behold, a year later, it had completely busted. So, as I was finishing college, I decided to go to grad school because I wanted to finish education. I did not want to stop. I’d seen both of my parents go back to college, both earning their college degrees. And I actually remember going to my dad’s master’s graduation when I was a young person, all of about twelve. So, I knew that education was important. That’s what they instilled in me. That no one could take education away from you is what I was told quite often as a child. And so, I wanted to finish. I wanted to be a doctor. My maiden name is Hill — I wanted to be Doctor Hill. I’ll get that over with so then I can decide what I want to do with my life.
CA: So, we’ll do our final question, which we we ask everyone as a way to end. For folks who really care about this issue of data equity, is there one takeaway or one resource that you recommend?
BM: So, when it comes to ‘data equity,’ I think the one resource would happen to be a number of books. I can actually share a number of books. I think some of these books have been in the media for a while now. But if you have not read them, read them. Algorithms of Oppression is Safiya Noble. Race After Technology is Ruha Benjamin. And Automating Inequality is Virginia Eubanks. Yes, they are three women — two Black women and one white woman — but I think this gets to gender, race, and class in three different perspectives. That is very important. So read those books, digest those books, have book clubs about them. That’s number one.
Number two would be to hire strategic ‘data equity’ firms and to really have a conversation about what your strategic goals are when it comes to your data practices. So, it’s not just about helping your organization earn money, it’s also about helping the retention of individuals, helping to understand what is the impact of your products moving forward. And also, it’s about, “Are you going to be a good company?” Or are you just going to be a company that earns money? So, I think the big takeaway is really start to think about the strategic vision, in the context of reading those books. And bringing that skill set forward to your workforce. And that does mean you’re going to have to spend money that’s going to be about educating your people. And it’s not going to be an automated system. It’s going to be small groups. It’s going to be collaborative. It’s going to be co-created with firms. Because that’s the only way, in order to bring humanity into this particular perspective.
CA: That’s great, and I think we can definitely make those book recommendations available to our community, along with other book recommendations from other folks we’re speaking to. I think we should make sure we’re circulating those and urge people, as you’re saying, not just to read them, but to talk about them.
Well, thank you so much, Dr. Marshall. This has been really fascinating and really invigorating. I have lots to think about and I’m sure everyone who is listening does too. So, thank you so much for joining us.
BM: Thank you so much, Colleen, David. This has been fantastic. Hopefully this helps folks and we can have a continued conversation and some action behind all of these conversations.
DH: Yes, thank you. Thank you very much for joining us today. That’s a wrap on our interview, but the conversation continues.
CA: Remember, we want to hear from you, so please send us your reactions, thoughts, ideas, comments to email@example.com.