[00:00:08] Lauren Burke: Welcome to Women in Analytics After Hours, the podcast where we hang out and learn with the WIA community. Each episode we sit down with women in the data and analytics space to talk about what they do, how they got there, where they found analytics along the way and more.
[00:00:32] Lauren Burke: I'm your host, Lauren Burke, and I'd like to thank you for joining us.
[00:00:38] Lauren Burke: So excited to have Dr. Brandeis Marshall joining us. So welcome, Dr. Marshall. Thank you for taking the time to join us and I'm really excited for the conversation we're going to have.
[00:00:48] Brandeis Marshall: I am so excited to be here as well, Lauren.
[00:00:51] Lauren Burke: To start off, could you tell us a little bit more about your background and the path you've taken over the years?
[00:00:57] Brandeis Marshall: Oh wow. So we only have a few minutes to have a conversation. So let me just start by saying that my road to data was actually forged when I was a child. And that was with rolling coins, as well as helping my father organize membership for an organization that he was volunteering with as part of his organization.
[00:01:19] Brandeis Marshall: So I got into data mainly because I loved math, but I thought math was boring. So I wanted to have a career and a life that included something mathematical, but did it with a creative twist. So that's where I found computer science. And so I saw code as a way to just do math models, but do it in a way that I could express how I want that code to look and how I want that code to behave.
[00:01:50] Brandeis Marshall: And I always was so concentrated, so concentrated on the impact of everything that I did. So anyway, let me, let me fast forward to analytics. When it came to analytics for me, I was so concerned with how data was modeled, I never even imagined just how much strife and complication was involved in analyzing the data. So I came to analytics really as this backend process of my curiosity of, "Okay, I've modeled the data well, now how is that data actually applied." Again, trying to make sure that I stay with my full theme of impact. How is what I'm doing impacting the world, the society of people who look like me? So that's how I came to analytics. It really was just me just trying to do good work.
[00:02:46] Lauren Burke: That's awesome. And I think that's a lot of what we see with people in data spaces. There's so much data. You can have all the data in the world, but if you can't take that data, transform it into something that helps you better understand the world around you, make some decisions off of that, then what are you gonna do with all of that data? It's just, it's just sitting there.
[00:03:05] Brandeis Marshall: Exactly, that was exactly my problem. Everyone was like, "okay, you were able to model the data, but then what happens after that?" And I was like, oh, good question. What does happen after that? What models are useful? What models are not useful? When are certain models actually should be applied? And when should other models be substitutes? What are alternatives? Right? So those are all those type of questions that bubbled up as I looked more on the analytics side of things.
[00:03:33] Lauren Burke: Right, and I'm sure the more you looked into it, the more questions come up. The more those all take you in more and more different directions.
[00:03:39] Brandeis Marshall: Yes. And then I continued down, like, the data pipeline. I was like, okay, now that I understand what's happening on the modeling side, now what's happening on the analytics side? And then it went to what's happening on the visualization side. And then lastly, what is happening when it comes to, you know, storytelling and what is happening when it comes to ethics.
[00:04:00] Brandeis Marshall: What is the implication in our society? Right? So it just all kind of bubbled and just kind of, I didn't wanna kick, I didn't wanna kick the can down the road, but I kept trying to figure out what is the impact as more and more data is being fed and manipulated and being fed into these different systems.
[00:04:18] Lauren Burke: Right, and the data ethics side of things is really what we're here to talk about today. Since back in September you published a book, Data Conscience: Algorithmic Siege on Our Humanity, which explores how data structures can help or hinder social equity. So what inspired you to write that book?
[00:04:36] Brandeis Marshall: What inspired me was the lack of this book. I, I was looking for a reference of how do you bridge this gap between. What is happening inside of the data world? What's happening inside of the data pipeline, right? How is the data life cycle being used, and then how is all of this data then being applied to society?
[00:05:03] Brandeis Marshall: What are the social, the historical, the economic, the political implications? And what are the ramifications of all this? Because I knew back in 2000 that data ran the world, right? Everyone's just caught up in the past couple of years.
[00:05:19] Brandeis Marshall: But for me, I was like, okay, there's still a lot of books about how to manipulate data. Based on the programming language, you can use SQL or SQL depending on where you stand on that and any of the other platforms and data tools. And then there's a whole lot of conversation about what is the societal impact of that data. But there was nothing in the middle. There was no bridge.
[00:05:44] Brandeis Marshall: And so I wrote the book to be the bridge, to literally be the bridge between these two worlds. The, the data processes are happening and they're happening at a fast rate. The impact of these processes are happening at a fast rate and affecting a lot of different types of people in different and sometimes harmful ways. And we need to have a book and a reference that bridges these two. That connects them so that we can humanize what's happening inside of our digital infrastructure.
[00:06:21] Lauren Burke: Right. And if you're thinking about the impact after it's already had that impact, right? Especially in places where you have adverse effects. That's too late. You need to be thinking about it when you are starting off, when you're talking about that first book, how you can manipulate your data, how you can set up your data practice.
[00:06:39] Lauren Burke: Those need to be something that's already in your mind. You're already kind of thinking about and planning to make sure that what you're doing aligns with the impact you want to have and the impact that you want to prevent.
[00:06:51] Brandeis Marshall: Exactly. I mean, this is what I was finding too often. Was that all of the ethical questions were happening after the product had already been created. Not even just in the design phase or just even in the planning phase. Like after it's already implemented, after it's already been deployed sometimes. Then people are asking questions about the ethics, about the social impact, about the societal impact.
[00:07:17] Brandeis Marshall: And I was like, hold up, we gotta back this train up. And go way back to the beginning and let's make sure to have the ethical conversations at the front end of the data process, not on the back end of the data product.
[00:07:34] Lauren Burke: That's a great point. Especially if you're thinking about it before you begin, instead of trying to later tailor your work to make it more ethical or to now abide by these accountability standards that you've set for yourself in the beginning, right? Then you already have that kind of as the framework, as the principles you're following along the way.
[00:07:53] Brandeis Marshall: Exactly, and a lot of the conversations and readings that I was doing was in the medical field or the healthcare realm as well as in the insurance and the banking realm, because those are two industries that have a high level of compliance, regulation and policy set up. I mean, they're, they're very regulated industries. There's a lot of oversight.
[00:08:18] Brandeis Marshall: So their notion of compliance and regulation and making sure that there is governance around their tech products is something that is pretty much ingrained inside of these industries. But that does not apply to the general environment when it comes to tech. Tech communities really don't have that standard.
[00:08:40] Lauren Burke: Right. And it's not something you should be told to think about. It should be something that you're already thinking about when you're doing things that could be potentially affecting a large amount of people.
[00:08:49] Brandeis Marshall: Exactly. Exactly. And, and nor is a lot of software developers or data professionals even taught. Like, it's not really part of the mainstay of curriculum. It's not part of the core curriculum in a lot of undergraduate, graduate, and even professional environments. I mean, you have to learn on the job. And so hopefully this book will be one where you can use it in a classroom, but you also can take it to your team meeting, right.
[00:09:19] Brandeis Marshall: It can be on the coffee table inside of the boardroom to say, Hey, maybe we need to look at some of these questions that I've raised and look at some of the recommendations that I've put together. And have these conversations at a, you know, more general level. And then how does that feed into practice inside the organization and how the organization runs its, its services and its products.
[00:09:44] Lauren Burke: It sounds like the audience for this book could be very broad, because a lot of people can benefit from the content. Who do you think it is most essential to learn from the perspectives shared within this book?
[00:09:56] Brandeis Marshall: I wrote the book for the junior data professional. Like the new data analyst, and hopefully the person who is mentoring or managing the junior data professional. That's who I wrote it for, because I figured those are the people that are gonna be in the trenches dealing with the data. And the managers are gonna be the one getting the problems from the junior data professional and trying to help that person through those challenges.
[00:10:29] Brandeis Marshall: But now that the book is out and I've received some feedback, I, I think the book is also intended for the managers to really start to think about how they engage the people who are in their direct report. What are some of the standards and operating practices around data inside their teams, and also inside the organizations, that they want to institute in order to make their lives a little bit easier as managers, but also to elevate the conversation so that there isn't this strife or friction. There isn't a lot of disagreement in how people will deal with data, right?
[00:11:14] Brandeis Marshall: So I think it's really two different audiences. But I'm very hopeful that those that are being taught right now, especially if they're graduate students that they will be able to see the book inside their classroom. Or upper level undergrads would be able to, to benefit from the book. Like some of the content is not really designed for the K12 audience. But I think if you are, you know, upper level college student and of course working, I, I think you'll, you'll find, find a home with this book. You will feel like, oh yeah, I, I remember having this issue, how do I deal with that?
[00:11:55] Lauren Burke: That sounds incredibly helpful. And like you said, a lot of programs in data analytics, data science often don't include those courses that focus on data ethics. Which I think is a huge opportunity as well. But for those people who are at that point and they haven't really been out in the industry or they're just starting out, I feel like this book could be a tremendous resource just to cover some of those gaps before they really are in the trenches, are needing that information to kind of backtrack some issues that might have sprung up.
[00:12:26] Brandeis Marshall: Exactly. Exactly. I mean, that's the reason why I separated the book into three sections, right? The first part is transparency, the second part is accountability, and the third part is governance. Because I think there needs to be a thought of how the technology is currently working, how it should be working, and then again, how do we reinforce some good data practices and making sure the data ops actually is helping humanity and not harming humanity.
[00:12:51] Lauren Burke: Right. And so what we're gonna kind of dive into today is the governance part of it, specifically around algorithms. And one of the things that came up while we were talking about this is the concept of algorithmic shadows. Could you tell us a little bit more about that?
[00:13:05] Brandeis Marshall: Yes. So this is a concept that is not mine. So first and foremost, this is a paper written by Tiffany Li and that's L I. She put together this wonderful concept called algorithmic shadows, which is essentially, you can delete the data, but if you do not also delete the algorithm, there will be data remnants inside of your system everywhere. So deletion is kind of a mirage, is essentially what she is talking about. Which I find to be absolutely brilliant. And, and this is one of the problems with the fact that data doesn't die. Data has a shelf life, but yet we don't put it on the shelf. So essentially algorithmic shadows is about the fact that data just keeps on living even when you try to quote "delete it".
[00:13:57] Lauren Burke: Right. That kinda leads us into the next topic, which is something you reference in data conscience, the relatively new concept of algorithmic destruction. So how does, how does that come into play with some of the new accountability laws we're seeing?
[00:14:12] Brandeis Marshall: Yeah. So a little bit of history, cuz I am a little bit of a history buff. There has been algorithmic accountability acts that were done in 2019 and also in 2022. In 2019, it was a very thin proposal that did not necessarily have a lot of conversation with technologists and data people, so it wasn't very practical or actionable.
[00:14:35] Brandeis Marshall: But they went back to the drawing board in the United States, and talked to technologists and data professionals and built up a wonderful slew of accountability measures. And one of them happened to include this notion of algorithmic destruction. And this is something that the FTC has actually implemented a handful of times.
[00:15:00] Brandeis Marshall: Algorithmic destruction is not only the penalty for malfeasance. Bad acting when it comes to a company, a corporation, and dealing with data and breaching any type of data, privacy of its customers or clientele. It also is removing the algorithm and the data that came from that algorithm. So it is almost as if there is a hard reset.
[00:15:27] Brandeis Marshall: Which I find to be a wonderful trifecta. There is a sanction, there is a penalty that tends to be monetary that has been happening for decades. But then there's this third component, which is removing the algorithm, removing the data that came from that algorithm that was causing that harm or found to cause that harm.
[00:15:49] Brandeis Marshall: I find it to be a fascinating application. And it addresses this issue of algorithmic shadows that Tiffany Li brought up in her paper. So this is how something that you think is just academic has now gone into practice, which to me is brilliant.
[00:16:08] Lauren Burke: Right, and there's already been examples of companies who have been found to violate policies that have them having to go through with destruction of their algorithms.
[00:16:17] Brandeis Marshall: Yeah, there's, there's several that has actually been implemented. And of course I cannot think of what they are right now. I completely have a brain freeze about that. But these companies have had heavy fines. Like millions of dollars in fines.
[00:16:32] Brandeis Marshall: But then the fact that they have to now get rid of that data, which means they cannot monetize, right? They can no longer monetize that data. Which as we know, data is a new currency. If you have the data, you're able to of course use it and put it in other systems and then be able to generate revenue.
[00:16:53] Brandeis Marshall: So, as I said, it needs to be used more often. I'm looking forward to seeing many more examples of the FTC and other organizations using algorithmic destruction as a penalty for companies that are breaching trust of us as citizens.
[00:17:14] Lauren Burke: I think it could also set a good example for companies to see that data, which right has a lot of value. Especially if you built a model on a very, very large data set. The penalty being you lose all of that data, you lose all the insights you can gain from that, is something that I feel like could definitely affect the way people are approaching building new algorithms, building new models, especially those that have the potential to harm people.
[00:17:38] Brandeis Marshall: Exactly, and also for those that are developing these new tools and these new algorithms. It also lets them know that not all the code is bad. But that bad code exists in conjunction with some of the not bad code. Because I think there's this notion within at least computing, and I would argue also the data world, that either code is all good or all bad. And I think we need to start to think about the gray area, right?
[00:18:12] Brandeis Marshall: Two plus two equals four. That's not a bad thing to code up . What is bad is when we reduce people to numbers and then the final number marginalizes them or amplifies them in not appropriate ways. So I think that contextualization is very important in order to make that distinction. So hopefully the tech world in general will start to move toward this understanding.
[00:18:41] Lauren Burke: Right. That's a really interesting point. I think you might also see where what started as good code, a good model, turns bad due to something like data shift. Or just now you come across a situation where someone was marginalized, someone was negatively affected, and then it changes the entire way you view that model.
[00:19:00] Brandeis Marshall: Exactly. I mean, especially in the, the lane of analytics, right? There's so many models that have been established and used quite frequently, but sometimes those models have been applied badly. And so it would be a great outcome to say, Hey, this model does not perform well and should not be used because it causes X, Y, and Z harms. So let's go ahead and impart algorithmic destruction on that. So then we remove not only the use of that algorithm, i.e., deprecating it. As we know what to do in code, we deprecate it. We can deprecate that particular use of that analytics model. And then we also can remove all of the data associated with that model. So we don't then spoil the rest of the data that's in the system, right? We're like removing the bad bruised apples, versus saying the whole bushel needs to be thrown away.
[00:20:05] Lauren Burke: Do you think there is possibility for some types of algorithms, that haven't been proven to accomplish what they say, say for detecting the emotions on faces, for those to be coded into law to, to be set in policy.
[00:20:21] Brandeis Marshall: I think we're gonna get to the point where any type of facial recognition, emotion detection, are going to be under a big spotlight. Because if someone has a natural face that is deemed to be aggressive or deemed to not be on the positive end of whatever that spectrum is. That is again, inducing harm.
[00:20:48] Brandeis Marshall: So we need to be cognizant of these things as developers and programmers and users, right, of these technologies and these innovations. I'm hoping the law will pass that will make them, there be bans. I know it will take some time.
[00:21:05] Brandeis Marshall: And that's what I talk about in the book is just how long the law takes. It takes like a hundred years to do like anything. I'm hoping it won't take that long, but sometimes I'm just not as optimistic. I think when I wrote that part in the book, I wasn't optimistic. But there's definitely occasion in where, where to be very cautious of the biometrics and, and hopefully our government will not be leaning into the biometrics.
[00:21:29] Brandeis Marshall: We'll see. There's been some facial recognition laws that have been reversed in the past, you know, few months. I think Virginia and New Orleans is the most recent. So I'm a little cautiously optimistic now, seeing those reversals.
[00:21:46] Lauren Burke: That's good to hear. I think some of the time we are so invested in how innovative some types of technology are that we um, sort of wanna hold on to that perfect world where it works and it has such an amazing impact. And then it's hard to look back and say, oh, this isn't really working as intended and the negatives don't really outweigh the benefit we're getting, and then you have to backtrack. Which is hard with some of these new really cool things we're doing. But I think it's good that it's being done, that people are taking accountability for that and working to change that.
[00:22:21] Brandeis Marshall: Yeah, I think there's a lot of people who are just asking about the nature of their work and wanna make sure that the work that they're doing in the digital sphere, that tends to be around code, you know, developing tools and platforms and systems and algorithms. That what they're building isn't going to induce any more harms then what's already in the system.
[00:22:47] Brandeis Marshall: And then there are slate of people who are trying to actively work to figure out what harms currently exist in trying to do their best in order to get rid of them. So I think with the these two groups working in tandem, we can get to a place that is much more equitable, which is the whole hope of the book, right?
[00:23:07] Brandeis Marshall: So the book is hopeful at the end of the day. Is that, hey, there's things you can do. There's things that you can do if you're a lay person, there's things you can do if you're in the tech world, things you can do if you're a data person. There is hope in order to make this better. But we have to call the thing
[00:23:24] Lauren Burke: Right. So including some of those groups of people, who else should be involved in the process of creating and introducing some of these future laws and policies that ensure companies and individuals are accountable for the algorithms they're producing?
[00:23:38] Brandeis Marshall: I mean, I think that there, there needs to really be community effort. I think the public. We as the public need to be more involved. I mean, I have a whole section talks about data civics and really talking about data from yourself. So data and you to data at home. So you and your immediate family. Data in the workplace, you and your coworkers, how do you share data to, of course, society different levels of society.
[00:24:06] Brandeis Marshall: And I think that's where we need more people to understand that data isn't a bad word. It's not the bad four letter work that you think of when you're a kid. And, and that data is much more than numbers. It is the audio, it is the images, it is the text, it is the smells, it is the feels. It is everything that surrounds us and how we can be comfortable with talking about content.
[00:24:30] Brandeis Marshall: If you wanna call it content, let's call it content. If you wanna call it information, let's call it information, but at the end of the day, it's all data. And more, more of us need to have the literacy around how our data lives are. I mean, we have these phones that we carry around and these phones.
[00:24:50] Brandeis Marshall: Carry a lot of data. It carries people's credit card information. It carries a roster of everyone that's in your contact list. It carries your mail, it carries what services that you engage in. So we are already part of the data industry. We are part of the data economy, and I think more people can enter in and contribute to putting pressure on the laws, putting pressure on companies in order to do right by us.
[00:25:22] Lauren Burke: So, other than the kind of time limitation. I think in your book you said you expect it to be two to three generations before some of these things will change. What are some of the other challenges and limitations in developing these policies?
[00:25:37] Brandeis Marshall: I think it's really knowledge. There's so many things happening in our lives. So much data's coming at us at one time. I had a number at one point of how much data that we ingest versus how much data we produce, and it is in the hundreds of megabytes a day that we generate, as well as ingest in a day.
[00:26:01] Lauren Burke: Wow.
[00:26:02] Brandeis Marshall: That's a day. So I think there's so much that is taking our attention away from how we are being perceived and how we are being used or abused inside of the data infrastructure. So we need, we need to be much more tuned to listen. Everyone's inside of the digital world. We need to pay just as much attention to the digital world as we do the physical world.
[00:26:37] Brandeis Marshall: Not that we completely say we're only digital, but just to understand that our data is floating around inside of lots of different systems, and we need to understand what that means for us, what it means for our children, what it means for our parents. Because multi-generations are being affected, people are being scammed, people are being doxed, identity theft, people are being socially engineered, right?
[00:27:03] Brandeis Marshall: So we need to be cognizant of these things and to make our world better. I know it sounds like a broken record, but our world can be better if we pay attention to what's important. And what's important isn't all of the celebrity news. What's important is making sure that your location services are turned off on your phone, making sure that your credit card information isn't on all the platform Like these are important tidbits to know and we need to have that public data education.
[00:27:39] Lauren Burke: Right. Just allowing people to know that how they, as an individual are affected by the data that they send out, the data that comes into them, and just where they're providing it whether they're opting into things that are using their data in ways that they're not even aware of.
[00:27:55] Brandeis Marshall: Right. It's that notion of when people realized with their mobile phones how many apps they had open. And their screen time, everyone started closing their apps, right? So I think it's the same type of acknowledgement as like. Oh wow, like I give out my data all the time cuz I'm logging into all these different separate systems and all these different separate systems have different pieces of who I am as an individual.
[00:28:26] Brandeis Marshall: So let's make sure to kind of, keep track of that a little bit better, a little bit more. In order to make sure that, you know, we can protect what we want to keep private as private.
[00:28:39] Lauren Burke: So earlier this summer, I learned about this concept called predictive privacy, which was introduced by Rainer Mühlhoff in a 2021 paper. And basically it comes down to that an individual, or a group's, predictive privacy is violated when sensitive information about them is either guessed or predicted by matching it with information from others. And so depending on how those predictions are used and shares, this could lead to unfair treatment or biased decisions against that individual.
[00:29:07] Lauren Burke: One example is like from Facebook data, it's possible to infer a person's relationship status, their sexual orientation, religion, their political views, behavior, or even health issues just based on how they interact with the platform. And for some people that could lead to really unfair decisions, right? Unfair treatment. So do you think um, individuals should have the right to predictive privacy?
[00:29:31] Brandeis Marshall: Yes. I had never heard that term. And you explained it. And I was like, yes, this is a similar term it's called data mining. With there's associations that can be made. So early days of data mining, there was a lot of results coming out, talking about how people tend to buy milk with bread.
[00:29:54] Brandeis Marshall: It's kind of the same notion. You can infer what people need based upon their buying habits. There also was a correlation between people who were buying Pampers or diapers and beer. a way that grocery stores were set up typically now have the diapers next to the beer idol. Why is that? Because typically, the partner of the person who birthed is the one who goes out to get the diapers, and that person also gets us case of beer.
[00:30:30] Brandeis Marshall: So predictive privacy reminds me of that type of notion of data mining. But yes, I think this is part of the issues that need to be brought to light for individuals to know that how they interact with these digital systems are providing data. Even if you're not providing the concrete yes, no indicator, you're providing a probability that is leaning a certain way. Probability, that's too big of a word, for most folks who are not in the data industry, but just to let them know like how you engage on a platform does share things about you.
[00:31:11] Lauren Burke: Right. I thought this was super interesting because it's against your will, right? Maybe you're opting in to share your data, but it's essentially saying things about you, predicting things about you based on your information to understand things about you that you probably wouldn't be willing to share with those platforms or those apps.
[00:31:30] Lauren Burke: So I think that's a really interesting new concept that's coming about. Just to say not only do you have the right to your data, to keep your data protected, but do you have the right to the data that you choose to share, to have that not be used to learn additional things about you that you might, might not want to share.
[00:31:48] Brandeis Marshall: Yeah, I mean, I say this quite a bit in one of the big, you know, data ethics issues is the fact that essentially two companies know everything about us. And I only say the two companies cuz everyone was like, oh yeah, yeah, yeah. That those two, I mean, it was three about 10 years ago, but now it's only two. Only need two companies and before you know, it's just gonna be one company pretty much knows everything about you.
[00:32:13] Brandeis Marshall: And that is alarming. And some people accept it and is okay with that. But other people are like, wait a second, I don't necessarily want these two companies to know everything about me. But yes, I think as a human right we should have the individual ability in order to say how we want our data to be used. But that of course means that companies have to be transparent in how they're actually using the data.
[00:32:38] Brandeis Marshall: And there's a lot of laws in place that are protecting companies from sharing what they were put under IP or sorry, intellectual property or trade secret. So how do we work around those too well establish ed tenants of the law in order to make sure that an individual person has a fighting chance in order to protect their data from organizations, to use their data and actually do predictive privacy.
[00:33:09] Lauren Burke: Yeah, there's definitely a lot more that can be done on the transparency side, just to make sure that as we're kind of coming into the knowledge that as individuals our data isn't our own. I feel like most people understand that by now, but they don't really understand the depth of that, like how far that goes. And I think that's, that's important as we continue to find more and more different ways to capture more and more different types of data, that that's also something that people are understanding just how, how far that can go.
[00:33:40] Brandeis Marshall: Yes, exactly. I think there is a term that is being used quite a bit, which is like creator economy, right? There is the creator economy. Well, that means people are creating content. And so if you're creating content, is that you as an individual creating content or are you bringing on other people to help you create the content?
[00:34:04] Brandeis Marshall: Right. So who then owns the content? Is it the person who's hosting, or is it the person who's bringing their value and their knowledge? It should be shared. But typically the way it works in creator economy is that the creator, whoever's the host, is the one who actually owns that content.
[00:34:25] Brandeis Marshall: Now we're getting into some interesting waters of conversation of what does that mean when it comes to creating? Who is actually the creator? Is it the person hosting or is it also the person who is, let's say, the guest or bringing in their value, sharing their views and things like that. So there's, there's a lot to unpack with just how we think about our data, what we consider data, what we're comfortable with sharing. There's a lot, there's just a lot. This conversation could go on, I think probably for like another two hours.
[00:34:57] Lauren Burke: Right.
[00:34:58] Brandeis Marshall: There's things that people don't think about. They don't think about how, you know, they type something into a platform and how the platform then uses it and gets a certain amount of money from it.
[00:35:08] Lauren Burke: Right.
[00:35:09] Brandeis Marshall: And how they don't get any money from it. The person who actually, you know, was the originator of that content. Doesn't get any money from it, but the company does.
[00:35:18] Brandeis Marshall: Right? Conversations happening on these big platforms about, okay, how much viewership do you have? And then per view, you get certain number of cents.
[00:35:28] Lauren Burke: Right, but the value of that trickles down from the host back down to the actual content creator,
[00:35:34] Brandeis Marshall: Yeah. And what does that look like? So there's a lot of conversations that I think will bubble up in the next few years. It's gonna be about how do we deal with all of this data? As I said before, we're each generating 150 megs of data a day.
[00:35:50] Lauren Burke: It's a ton of data.
[00:35:52] Brandeis Marshall: A lot of data. You know, that's a lot of data.
[00:35:56] Lauren Burke: Well, kind of speaking of content, are there any ongoing or future projects that you're excited about?
[00:36:04] Brandeis Marshall: Oh my God. There's so many projects that I'm working on. So recently I hosted the first, the inaugural Black Women in Data Summit. So I'll be planning another one in 2023. So I'm gearing up to start that planning at the early part of January.
[00:36:23] Brandeis Marshall: There's also a number of posts that I do per month on Medium, so you can check me out on Medium. I write about, you know, data ethics and just the data industry in general. On there.
[00:36:35] Brandeis Marshall: And then of course, I'm just posting on social media. I am figuring out LinkedIn versus Twitter. As we've seen, you know, Twitter go through its metamorphosis, I'll call it that. I'm very kind by calling it that.
[00:36:54] Lauren Burke: It's an interesting time on Twitter. We're all aware of it.
[00:36:58] Brandeis Marshall: We're all aware. But yeah, so those are the main things that I'm thinking through and just trying to engage people in. And just in general, my company, DataedX Group, we're just open for projects to do interesting things, right?
[00:37:10] Brandeis Marshall: Need someone to curate, want someone to do some data research. I'm here for it. I have, there's several interesting projects kind of in the work on the, you know, learning and development side. Just getting more people to understand and be comfortable with data and dealing with that tension of some of the data ethics issues.
[00:37:28] Brandeis Marshall: Yeah, just lots of different things are happening cuz people are having these conversations.
[00:37:33] Lauren Burke: That's awesome. So for those, just to plug, for those interested in working with you at DataedX, what's a example of a project you've worked on that you are really excited about?
[00:37:44] Brandeis Marshall: So example of a project. A recent project that I worked on and you can go to it, it's datacareerpaths.com. And it was a project in which I brought, with collaborators, a research project that was all about looking at what is the likelihood of certain demographic groups being in a data classroom as a grad student in the United States.
[00:38:11] Lauren Burke: Interesting.
[00:38:12] Brandeis Marshall: So as I said, it was a research project I was working on with Dr. Thema Monroe- White and a couple other folks. And then we took that work and then we visualized it. So you can go on and see Tableau visualizations. So we looked at the cross section of race, ethnicity, and gender. So we looked at that intersection.
[00:38:31] Brandeis Marshall: So you can see if you're a black woman, where would you see yourself? What institutions that actually have data programs. And what's that likelihood. And so just to give a little tease, so for black women there's a 2% chance that you're gonna see yourself in a data classroom.
[00:38:49] Lauren Burke: Whoa.
[00:38:51] Brandeis Marshall: 2%. And that's if you're a student. It's about 1% if you are an instructor.
[00:38:56] Lauren Burke: Wow.
[00:38:58] Brandeis Marshall: And so we break down institution by institution. You can see the scale. There is, you know, overrepresentation, underrepresentation, scale that we put inside the visualization. But we also provide some resources for where you might want to go in order to look up data programs, look at the schools themselves cuz you might be location bound and so you wanna find places within your state or within your community and we kind of help you do that as well.
[00:39:25] Brandeis Marshall: So that's a really cool project that was released earlier this year that I'm super proud of and it's just a very cool project.
[00:39:32] Lauren Burke: Yeah, it sounds like it could be a really, really valuable resource. So we'll make sure to link that so everybody can check it out.
[00:39:38] Brandeis Marshall: Yeah, we don't have any more funding, but we would love to continue the project and get more schools with more data. In order to kind of make the data like really concrete or really accurate, right? Because we had to use public available data, which we know may not be the most accurate.
[00:39:54] Lauren Burke: Right. So speaking of resources, I like to ask everyone. What is a resource that's helped you in your career that you think might help others who are listening?
[00:40:02] Brandeis Marshall: I think the resource is going to be AI Ethics Weekly, which is a newsletter that comes out every week. And it talks about the AI ethics world. It's a great resource. It is a nice little bit of what is happening in the data ethics world, what is burning down, what is helpful. They also have, of course, a jobs board, as well. So if you're interested in getting into the field, there is a jobs board as part of that newsletter, but it comes out weekly and I think it's fantastic.
[00:40:36] Brandeis Marshall: Outside of my own newsletter, of course. I have a newsletter called Rebel Tech, and so Rebel Tech Newsletter is another resource as well.
[00:40:45] Lauren Burke: Nice. Well, we'll definitely add that as well. Those both sound like they will be very helpful. Um, but how can our listeners keep up with you?
[00:40:53] Brandeis Marshall: Definitely go to LinkedIn, so it's just the LinkedIn type in Brandeis Marshall. You should be able to find me. There's not too many Brandeis Marshalls in the world especially with my picture. So check me out there on LinkedIn. And then of course I can contact me on my website. Just brandeismarshall.com.
[00:41:10] Lauren Burke: Awesome. Well, thank you Dr. Marshall. Thank you so much for joining us today. I really enjoyed talking with you. And I feel like I learned a lot and I'm excited for others to be able to listen and learn from you as well.
[00:41:21] Brandeis Marshall: It's wonderful to be here. Thanks for having me. This was fun.
.png)
