Kirill Eremenko: 00:00:00
This is episode number 407 with Co-founder and Co-director of Women in Data Science, Margot Gerritsen.
Kirill Eremenko: 00:00:12
Welcome to the SuperDataScience podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur, and each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
Kirill Eremenko: 00:00:37
Welcome back to the SuperDataScience podcast everybody, super pumped to have you back here on the show. Today we’ve got a very special guest calling in from California, Margot Gerritsen. What you need to know about Margot is that she is a professor at Stanford, a full professor at Stanford, she’s been there for many years, and she teaches some very interesting topics in the spaces of mathematics, computing, and energy.
Kirill Eremenko: 00:01:08
In addition to that, Margot is a co-founder and co-director at one of the most renowned conferences in the space of data science called Women in Data Science. If you haven’t heard of them before, check them out at widsconference.org. This is a conference that aims to promote diversity and inclusion and empower women in the space of data science. So a very, very noble cause. And in today’s podcast you will hear about quite a lot of very interesting things. So we’ll start off by talking about some travels, including how Margot went to Antarctica with her son on his 11th birthday, then we dive into linear algebra out of all things and you will get an overview of what linear algebra is, we’ll talk a bit about R versus Python, we’ll talk about principal component analysis or PCA and redundancy in data. In this episode, for the first time on this podcast, you will get a very intuitive explanation of principal component analysis, so don’t miss that. I loved Margot’s explanation, it helped me understand the concept very well.
Kirill Eremenko: 00:02:23
Also you will hear about some use cases, two different applications of principal component analysis, what it is actually used for. Then we dive into talking about Women in Data Science, the conference that Margot has co-created and now is co-directing. You will find out why Margot is so passionate about helping women in the space of data science and the myths that prevent women entering the space and thriving in the space, and what you can do to help no matter your gender or your background, there’s steps you can take to succeed yourself and to help others in the space succeed even further.
Kirill Eremenko: 00:03:04
So there we go, that’s what we’ve got coming up for you just now, very exciting podcast, can’t wait for you to hear it and without further ado, I bring to you Margot Gerritsen, Stanford professor and co-founder and co-director of Women in Data Science.
Kirill Eremenko: 00:03:25
Welcome back to SuperDataScience podcast everybody, super excited to have you back on the show, and today we’ve got a very special guest, Margot Gerritsen calling in from California. How are you going Margot?
Margot Gerritsen: 00:03:36
I’m doing well, Kirill. Thanks for having me on.
Kirill Eremenko: 00:03:40
Oh, thank you, thank you for coming on and for pronouncing my name correctly.
Margot Gerritsen: 00:03:45
I try, I try.
Kirill Eremenko: 00:03:46
The original pronunciation. It’s great to have you and although it’s unfortunate, you mentioned you were evacuated right now, right? You’re in an evacuation-
Margot Gerritsen: 00:03:58
Well, we’re back but we were evacuated. Yeah, the fires here this year have been tremendously bad.
Kirill Eremenko: 00:04:04
Okay. It started a bit earlier than last year, no? Like, last year was in December.
Margot Gerritsen: 00:04:15
We’ve had later and later actually in the year these wildlife fires come, so October/November is not unusual now and that’s the scary part because now it started the 16th of August here with these big fires because of a huge lightning storm that we had coming through central California and we’ve had these fires now for months, but we still have probably the worst to come so it’s terrible. We have over… I don’t know how many square miles have been burned, I don’t even want to think about it, it’s terrible. It’s the worst year, it’s the worst year. So right now the air quality outside is really bad, if the AQI means anything to you, we’re at 300 here which is hazardous.
Kirill Eremenko: 00:05:05
Wow, okay.
Margot Gerritsen: 00:05:05
But other than that California is still amazing. We still like the state but it’s sad.
Kirill Eremenko: 00:05:14
Yeah. Originally you’re from the Netherlands. What’s the difference between… And pardon my ignorance if this may be the incorrect way of saying it, Netherlands versus Holland? When do you use which word?
Margot Gerritsen: 00:05:29
I use both, I’m not so sensitive to that. But there are some touchies that really want you to say the Netherlands, and the reason is that Holland really is only two of our 11 I think provinces, so the country is divided in provinces and there’s two north and south Holland that really are sort of the center of the country in terms of the economy and most people live there and so on. But of course the other provinces don’t like it when you refer to the whole country as just those two. Most people don’t know, right? What the difference is. So I’m not sensitive.
Margot Gerritsen: 00:06:09
And we’re big soccer fans and we talk about Holland, the big cheer that we have for soccer is [foreign language 00:06:22] which means come on Holland, come on. So we don’t use the Netherlands. And Holland is a lot shorter, right, than the Netherlands that is a little bit easier to roll off the tongue.
Kirill Eremenko: 00:06:36
Yeah. Orange, right, is the national color?
Margot Gerritsen: 00:06:40
That’s right. Red white and blue and orange.
Kirill Eremenko: 00:06:43
Okay. And you still have a monarchy in Holland?
Margot Gerritsen: 00:06:46
We do, we have a king.
Kirill Eremenko: 00:06:49
Wow. It’s unusual to think about it, you usually just think about UK has a queen but it’s interesting that some, like Spain also has a monarch still, Monaco.
Margot Gerritsen: 00:06:58
Belgium.
Kirill Eremenko: 00:07:02
Belgium, wow.
Margot Gerritsen: 00:07:03
[crosstalk 00:07:03] and then of course some of the Scandinavian countries, so there’s still quite a few. But I don’t think in Holland the king and then formally the queen played such a big role, I never really thought about it as a kingdom in that sense. I’m not a big fan, but maybe I should go there.
Kirill Eremenko: 00:07:25
It’s interesting. Okay, all right, awesome. And you had a very interesting life, just from what we discussed before the start of the podcast briefly, like you lived in New Zealand for five years and now you live in California, originally you’re from the Netherlands. Tell us a bit about how did you end up being all over the world like that?
Margot Gerritsen: 00:07:49
By accident.
Kirill Eremenko: 00:07:52
Yeah? Wow.
Margot Gerritsen: 00:07:52
It wasn’t really planned. I often tell my students that the only thing that was planned in my life was that, from a very young age on, I wanted to leave the country. I think I was eight when I wrote in my diary, “I don’t want to stay here, I want to go west.” Maybe because I was a big fan of Led Zeppelin and they had the song Going to California, right?
Margot Gerritsen: 00:08:16
Then I remember when I was pretty young I watched the movie The Graduate, and I saw Dustin Hoffman drive across the Bay Bridge, which I actually thought was the Golden Gate Bridge, and I thought I want to do this, drive across that amazing iconic bridge in a convertible. So that got me going. So both of them, I thought California would be amazing, but I really just wanted to leave.
Margot Gerritsen: 00:08:39
If you’ve spent any time in northwestern Europe in the winter you know why because it’s gray and it’s depressing and Holland is very flat, I wanted to see mountains and I wanted to have more sunshine. So I left as soon as it was possible, which was really after my studies, and I got a scholarship to go somewhere in the world, and I decided to go to the opposite of Holland, which was Colorado, because the number of days of sunshine was about the same as the number of days of cloudiness in Holland. And they’re big mountains and it’s not flat, so I thought okay, this is it, this is what I have to experience.
Margot Gerritsen: 00:09:26
Then after a year there I got into Stanford, I applied to Stanford and I moved to California. Then got this job in New Zealand and I thought oh, let’s try that out and absolutely loved being there for five years. Then got an offer to come back to Stanford, which was totally unexpected. So I just jumped on it. So yeah, if you told me as an 18 year old or so hey you’re going to end up being on a podcast run by somebody in Australia while living in California, I would have said you’re absolutely out of your mind. I’ve been so lucky.
Kirill Eremenko: 00:10:03
Your world has changed so much.
Margot Gerritsen: 00:10:04
Yeah, also, yeah.
Kirill Eremenko: 00:10:06
It’s interesting. Lovely places, Colorado, California and New Zealand. It’s fantastic, fantastic that you’ve experienced these. What was from all your travels, what is your most memorable moment?
Margot Gerritsen: 00:10:29
Of all my travels?
Kirill Eremenko: 00:10:33
Yeah.
Margot Gerritsen: 00:10:33
Well, the best travel for me ever was going to Antarctica. I never lived there but I spent some time there, and the reason why this was so fun for me is I have a son, he’s almost 21 now but he was born in the year 2000 on January 11th, and I said to him when he was about seven, “You’re going to have one really, really special birthday because you’re going to be turning 11 on January 11 in 2011.” I thought that was kind of special. So when he was seven I said to him, “What do you want to do on this really, really special birthday?”
Margot Gerritsen: 00:11:13
And he was a big fan of penguins at the time, so he said, “Mom, I want to celebrate my birthday with a lot of penguins.” So I said, we traveled a lot my son and I, I was a single mom for some time, and I said to him, “You know what, I’m going to take you to celebrate your birthday with 25,000 penguins on January 11, 2011.” So I started saving up for a trip to Antarctica and that’s where we went. On his birthday he was on land in Antarctica amidst a penguin colony of at least 25,000 penguins with a cake and we celebrated his 11th birthday.
Kirill Eremenko: 00:11:55
Wow, that’s the most amazing story I’ve heard about travel or penguins. That’s so cool, you’re so cool.
Margot Gerritsen: 00:12:07
Well, that was fun. And the good thing is to do this when the kid is a little bit older so they actually remember. Anything that we did before the age of six he will have forgotten but I’m sure this he will remember.
Kirill Eremenko: 00:12:21
Wow, that’s awesome.
Kirill Eremenko: 00:12:27
Hey everybody, hope you’re enjoying this amazing episode, and we’ve got a quick announcement and we’ll get straight back to it. The announcement is that DataScienceGO Virtual number two is in town. It’s happening on October 24th/25th this year, and you can get your tickets today at datasciencego.com/virtual. And the best part? It’s absolutely free. We’ve got some amazing speakers, amazing workshops for you to attend, and of course the super cool part is we’ve got networking. There’ll be several 30 minute speed networking sessions where for three minutes you connect with a random data scientist from another part of the world or maybe from your part of the world, you get to chat for three minutes, if you like each other, if you want to connect, you hit the connect button, you stay in touch.
Kirill Eremenko: 00:13:12
This was by far one of the top features of DataScienceGO Virtual number one, so many people got such great connections, stayed in touch and some crazy stories came out of that. We’re going to repeat it and we want you to connect with your fellow data scientists. Once again, it’s absolutely free, register for your ticket today at datasciencego.com/virtual and I’ll see you there. Now let’s get back to this episode.
Kirill Eremenko: 00:13:37
Okay, fantastic. Well, when you’re not a mom, you are a professor and a senior associate Dean at Stanford. Tell us a bit about that, that sounds like quite a responsible important position at one of the world’s top universities. Tell us what you do in Stanford?
Margot Gerritsen: 00:14:00
Yes, very, very important. Well, it’s mostly just a lot of fun, and in fact I just stepped down as senior associate Dean. I’ve been really lucky that for about 12 years or so I was able to have some leadership positions on campus and my favorite one probably was directing an institute of computational mathematics at Stanford University, and I liked it particularly because I could work with all of the graduate students, and that’s the biggest joy. That’s why I’m at university is to work with young students. They keep me young and they’re idealistic and they’re excited and they’re super smart. It’s just a dream to work with them and to help them on their way, it’s one of the big privileges in my life, and I love teaching so that’s been great fun.
Margot Gerritsen: 00:14:54
And yeah, associate Dean, we had a great team of people working on what we call educational affairs. So again, really thinking about the teaching experience and the learning experience in our school. And Stanford, it’s a great place. It’s also, as you said, it’s a good university and so they expect quite a bit from you.
Margot Gerritsen: 00:15:19
My life now, I’m full professor so that means that you’ve been through every single promotion and my life now is a lot nicer than it was when I first started my career because then there is a lot of pressure on you and there’s a lot of… for most people I think a stress associated with that. But now I’ve got the life of Riley, right, as they say. It doesn’t get any better than this.
Kirill Eremenko: 00:15:45
That’s awesome, that’s awesome. And what subject do you teach as full professor?
Margot Gerritsen: 00:15:54
I really love teaching computational mathematics, and my favorite course I’m actually starting on Monday, that’s linear algebra and a lot of people think, well, why that? But it’s really the building block of almost everything that people do, and of course for data science anything that you do computationally, linear algebra really is the foundation, and so that’s why I love it so much and it’s a wonderful course that I’m teaching for beginning graduate students. So it’s wonderful to show them that even with only one quarter of this material they can do so much with that, they can understand principal component analysis, they of course understand regression. There is so much that they can do with this.
Margot Gerritsen: 00:16:47
So that’s probably my very favorite class, but I also teach classes in energy systems, particularly renewable energy. I’m a big fan of fluid dynamics, I did that for a long time, simulation of fluid flows and really excited about wind turbines. That must be my Dutch heritage. I mean we have had wind turbines for centuries, and we have a windmill in the village where I grew up, and I always thought they were beautiful machines. So I like teaching about that and explaining how they work.
Margot Gerritsen: 00:17:23
So that’s a lot of fun, and I’ve been teaching some courses in my distant past in fluid mechanics, which I also really, really loved. Yeah, a little bit of everything. I try to teach new courses frequently because when you’re teaching something you’re really learning the material, so sometimes when I think, “I don’t really know this well enough or that well enough,” I just volunteer to teach and then the shame factor kicks in and I’ve got to really understand it faster than my students. So I’m on this learning curve and I just find that very exciting.
Kirill Eremenko: 00:18:00
Gotcha. For someone who is listening and maybe is coming into the field of data science from a different field, less technical field, how would you caracterize or how would you explain what linear algebra is? How does it differ to just, I don’t know, calculus for example? What is linear algebra?
Margot Gerritsen: 00:18:23
You know, it really depends on what area they come in from. So the interesting thing about linear algebra, as I said, it is the foundation of many, many different areas in engineering and also the sciences, but for most people who are looking into data science, linear algebra comes in because data is typically stored in things like tables that we call matrices and then in data science you’re trying to understand the richness of the dataset that you have, and usually that means interrogating these matrices, trying to find out what is the relevant information in this matrix, what is redundant information in this matrix, and so you need to understand how to manipulate that. And linear algebra basically helps you do that. So you can look at a matrix as a set of columns or set of rows, and understanding how columns relate to each other, how rows relate to each other.
Margot Gerritsen: 00:19:39
So this is how a lot of people come in, and they want to do, for example, as I mentioned earlier, principal component analysis, which is really just a form of data mining. Where is the relevant information in the data that I’ve just stored? And you can translate that in a matrix manipulation or an algorithm. And so I teach these base algorithms that allow you to do that.
Margot Gerritsen: 00:20:06
Other people come into linear algebra from, for example, fluid mechanics or climate modeling or seismic analysis, and then matrices come up because they represent systems of equations and so maybe people who remember high school algebra remember systems of equations, but there they were maybe 2×2 so you had two equations and two unknowns.
Margot Gerritsen: 00:20:35
But in many of the systems that I’ve worked in, you may have millions of equations and millions of unknowns and you store them typically using matrices, and again then you want to be able to manipulate these matrices to understand these systems a little bit better.
Margot Gerritsen: 00:20:51
So that’s the interesting thing with linear algebra. I see this very much as the field that allows you to manipulate and work with matrices and vectors, and these matrices and vectors are used to store data, to represent systems, and therefore they come in so handy in many different places.
Kirill Eremenko: 00:21:15
Wow. Fantastic.
Margot Gerritsen: 00:21:17
I don’t know if that makes sense, Kirill?
Kirill Eremenko: 00:21:19
Makes sense, makes sense.
Margot Gerritsen: 00:21:21
A little bit, yeah.
Kirill Eremenko: 00:21:22
Yeah, no, it’s a great example, and it reminds me of the difference between R and Python, that R was designed originally to be very favorable to work with vectors and matrices, whereas in Python is feels like more of an extra, additional layer. That explains why R is so popular in science whereas Python is more popular among data scientists who came from the developer world.
Margot Gerritsen: 00:21:51
Yes. And of course now we see also a lot of students that come in from other areas of science using MATLAB, which stands for matrix laboratory, for that same reason. And Cleve Moler and other people started MATLAB because they understood that matrices are the building blocks of so much of science, so if you develop an architecture, software architecture platform that allows easy manipulation with matrices then you’re on to something, and basically R is based on that same philosophy.
Margot Gerritsen: 00:22:30
But I see a lot of the students, of course, liking Python particularly because of the visualization aspect of it, I think that’s often what they say, and also Julia, I don’t know if you’ve played with Julia, but that’s becoming really popular now too.
Kirill Eremenko: 00:22:45
I haven’t played myself but I’ve heard it is picking up quite a bit.
Margot Gerritsen: 00:22:49
Yeah. I haven’t myself either. We offer an introduction to Julia course and I took that once, but I let my students program in this, I do very little programming myself anymore.
Kirill Eremenko: 00:23:02
Gotcha. You mentioned principal component analysis, I would love if you could explain to our listeners, and to me frankly, what it is like just intuitively, what is principal component analysis and how can somebody just think about it to understand it better?
Margot Gerritsen: 00:23:21
Yeah, let me see if I can do this. So the name sort of gives it away. The name sort of says what are the principal components, in other words that are the principal pieces of information that you have available to you? So suppose that you’re storing a lot of data. Let me say that at every second in time, just as an example, you’re storing temperature data across the state of California, or every hour. And for every slice of that in time, so every hour or every second, you have all this temperature data and you put them in one big vector.
Margot Gerritsen: 00:24:14
So you just put them in one column of a table. So now you’re building this table for a whole year, every column represents a temperature distribution in California at a particular time. Then you start to wonder what is common about this temperature distribution? So now I’ve got I don’t know how many different columns, every column is a temperature distribution, but do they have a lot in common? So these columns, are there maybe some columns that are representative of the general type of distribution that I see? In other words, where does the principal content or information of this table really sit?
Margot Gerritsen: 00:24:59
Can I say hey, you know what, I don’t have to store that whole table. If I save columns 11 and 340 and 2057, then those three columns contain really the most important information. They can be used to represent the temperature distribution over the year, because you know what, from day-to-day maybe some of this changes but there may be three sort of temperature distributions that represent the yearly fluctuation of the temperature in California.
Margot Gerritsen: 00:25:35
So it makes sense that when you’re collecting a whole lot of data that there is a lot of redundancy in that data. So then you start saying, okay, if there’s a lot of redundancy in data, where are the principal pieces of information, where do they sit? So in principal component analysis, that’s what you’re really trying to do, you’re trying to distill, extract what are the most important pieces of information, that if people said okay, you’ve got this huge amount of data, but which pieces are really representative of the whole for the most part?
Margot Gerritsen: 00:26:12
Of course, you’re missing details, there are always exceptions, there are always outliers, but say for 95% you want to represent this data reasonably well, how much information would you need to keep to represent most of the behavior that you’re seeing? That would be principal component analysis.
Kirill Eremenko: 00:26:33
Wow. That is a beautiful explanation. Loved it. Fantastic.
Margot Gerritsen: 00:26:40
So then in this example that I gave you with the columns, finding out which columns you need to keep to be able to represent the data well, that would be PCA or in linear algebra we call it singular value decomposition, and it’s a wonderful way to look at… wonderful algorithm to have. And I grew up at Stanford, I did my PhD at Stanford in a group that had a guru of that principal component analysis. We call him the god of matrix computations, his name is Gene Golub, he came from Russia, or a Russian background, right, I think. Yeah, Golub?
Kirill Eremenko: 00:27:22
Golub, yeah.
Margot Gerritsen: 00:27:26
Doesn’t that mean dove? Does it mean dove?
Kirill Eremenko: 00:27:29
Yeah, I think it means pigeon.
Margot Gerritsen: 00:27:31
Oh pigeon, that’s right. Anyway, so Gene Golub, and so his ancestry was from Russia, I think he was born in the United States. But anyway, he was this unbelievable scientist and engineer, and matrix computation was his bread and butter, and he was the one who really created the efficient way to compute this singular value decomposition or PCA, and on his license plate on his car, when we were students, he had the license plate that said Prof SVD, singular value decomposition. So I’m a big fan of that algorithm.
Kirill Eremenko: 00:28:14
Awesome. Well, thank you for explaining it. I guess an analogy would be if you have a book and you don’t want to read the whole book, you just read the summary of a book. Instead of 200 pages, you read two.
Margot Gerritsen: 00:28:24
That’s right. And a lot of people make use of that. So they’re sort of making use of the principal component analysis of these difficult novels.
Kirill Eremenko: 00:28:37
That’s right. There’s even an app, I think it’s called Blinkist. It’s an app that you can download summaries of books, like any book they summarize and you just read the summary if you don’t have time.
Margot Gerritsen: 00:28:48
Yeah. And it’s a wonderful area of research, also, right? So if you ever had any text, can you automate the retrieval, if you want, of the most important pieces of this plot? That is an incredibly difficult thing to do of course, because that’s fascinating research.
Kirill Eremenko: 00:29:12
Yeah, well I think we’re getting there slowly with all the transformers and different types of neural architectures or NLP algorithms. So we’ll get there, I think.
Margot Gerritsen: 00:29:24
Yeah, that natural language processing is a wonderful area of research.
Kirill Eremenko: 00:29:29
Yeah. So one more thing on PCA, is my understanding correct that it doesn’t have to be identical data like temperature in these columns? For instance it could be the data of a store and we could have GEO demographic data about our customers, their purchasing habits, what items they like, all of this just randomly different types of columns, basically we could just normalize them and then apply PCA anyway, is that correct?
Margot Gerritsen: 00:29:56
Yes. And then what you often find is that you extract a bit of information from all of these different dimensions. So you have all of these different descriptors as you say represented in the data. Then when you start looking at this you probably extract a bit of everything, because-
Kirill Eremenko: 00:30:16
Gotcha.
Margot Gerritsen: 00:30:17
Yeah, for sure.
Kirill Eremenko: 00:30:21
And that helps when you build your model, whether it’s a logistic regression or a deep learning neural network, it helps you limit the amount of inputs you give it. So it speeds up the training and just makes it more efficient that way.
Margot Gerritsen: 00:30:35
That’s absolutely one of the ways that you can use it, you can just do data mining, where you’re only interested or primarily interested in the richness of the contents that you have, really trying to explore the data, what’s really in it, what’s not in it, what are the… like I said, the principal behaviors that I’m seeing here, the principal components.
Margot Gerritsen: 00:31:03
But you could also use it as an assistance in other algorithms and that’s often, as you say, really well said, to speed up other algorithms. Because you can imagine when you’re creating a deep learning algorithm and you have an enormous amount of data to train this, well you really only need to train it on relevant data, right? If you have repeated data, a lot of redundancy in the data, well that would be really nice if you understand where that redundancy is so you don’t need to train on the same thing over and over again.
Kirill Eremenko: 00:31:41
Yeah, absolutely. Well, Margot thank you very much for the description of PCA, that was I think the first time it was described so well on the podcast, so that will be very useful.
Margot Gerritsen: 00:31:52
You’re testing all your guests and saying, “Explain to me what is PCA?”
Kirill Eremenko: 00:31:58
No, no, no. I ask for something that feels right in the moment. But what I wanted to say is that now we’ve learnt a bit about your travel history and your teaching, so when you’re not being a mom and when you’re not teaching at Stanford, you’re the co-founder and co-director of WiDS, or Women in Data Science, which is a huge conference. How do you find the time? Like this sounds already like a lot of things. How can you, in addition to all of that, also be running a massive world renowned conference? I think you’ve reached over 100,000 people worldwide in 60 different countries. How do you find the time for all of this?
Margot Gerritsen: 00:32:40
Yeah, you know the secret is a really, really good team. And delegation, delegation, delegation. So WiDS, Women in Data Science, was formed almost… it was sort of a miracle, really, how that came together. We never set out with WiDS to build this huge, global conference, nor did I ever dream that this would become so big.
Margot Gerritsen: 00:33:17
But here’s what happened, in 2015 I got really sick and tired of having conference after conference in data science and when I talk about data science, when we talk about data science and WiDS, we see data science as this umbrella field that contains artificial intelligence and machine learning and data visualization and everything, right? And there is not a common, shared vocabulary.
Margot Gerritsen: 00:33:47
So some people see data science as totally separate from AI, so let me just make clear that we see ourselves as a container for everything related to data, including AI. But at that time in what I would call the field of data science, the most conferences particularly in Silicon Valley, I have to say, were totally male-dominated. We had so many conferences that only had male speakers, and I knew as a woman in this field myself, so many outstanding women doing really fantastic stuff and I didn’t see them as speakers at conferences or on panels, we have these so-called manels, the male-only panels. And said that’s not-
Kirill Eremenko: 00:34:38
I haven’t heard that before.
Margot Gerritsen: 00:34:38
No, the manels? Oh yeah, this is a common thing. So I got really sick and tired of it, and for me the last straw was a conference that was actually organized, I have to admit, on campus. And I had been asked to talk and I couldn’t make it on that day, and a little bit later I saw the announcement of the conference and I saw there were only really male panel speakers, and so I ran into one of the organizers and I said, “What happened?” He said, “Well, Margot, you couldn’t make it.” I said, “Are you kidding me, I’m the only woman?” Said, “No, well Fei-Fei couldn’t make it either.” I’m talking about Fei-Fei Li here, very famous AI person on campus, and she’s a wonderful colleague. “And we really looked for other women but we couldn’t find any.”
Margot Gerritsen: 00:35:27
And I thought okay, that’s the problem. They’re just not known, they’re out there, they’re not being promoted for whatever reason. Let’s change this once and for all. Let’s organize a conference that is a technical conference and just happens to have only female speakers, like so many conference at the time happened to have only male speakers.
Margot Gerritsen: 00:35:49
So I jokingly said the first conference we held when people said, “Why do you only have female speakers?” I said, “Well, Joe couldn’t make it. We asked him, but he couldn’t make it. And we really tried to find male speakers, but we just couldn’t find any.” So we started this and we of course invited men and women to come, but to our huge surprise we hit such a nerve and within a couple of weeks at the time for our first conference we were sold out and then the next year we thought, well maybe we should have this in more places, not just have a Stanford conference because we were of course limited the number of participants.
Margot Gerritsen: 00:36:31
We ended up, just because we thought it may be a fun thing to try, live streaming at the time. So this is pre-COVID, 2015, live streaming was not so common for conferences, and without ever-
Kirill Eremenko: 00:36:47
You were the pioneers.
Margot Gerritsen: 00:36:49
Well, we just tried it, and of course people had done some of it, but we started live streaming and we made it available to everybody for free, and to our big surprise we had 6000 people on the livestream without advertising. So to us, Kirill, at the time it meant we’re hitting a nerve. People and particularly women and young girls, they really want to see other women talking about this amazing work that they’re doing.
Margot Gerritsen: 00:37:19
And this was not a conference where we got together to complain about how hard it was for women in data science, this was a conference celebrating this really outstanding work done by all these amazing women. So we thought, well how can we scale this up? And Grace Hopper is a conference organized for girls and women in STEM, and we thought well we could do another Grace Hopper where we just go to a big convention center and get thousands of people to come, but to me that didn’t make any sense, and I thought well why don’t we use the livestream and make our lives a little bit simpler because as you’re saying if you have a really big conference it’s an unbelievable logistical nightmare and it costs a lot of time.
Margot Gerritsen: 00:38:06
So we were thinking instead of building a huge conference locally, let’s distribute this conference. We offer our livestream to anybody in the world who wants to use it, but around our livestream they can build a local conference. So my thought was, well, somebody in Texas could start a WiDS conference, they could maybe chime into the WiDS conference, dial into it from the livestream for a couple of the talks that we provide, and then the rest of the day they can have their local speakers. So we sort of spread the love a little bit more.
Margot Gerritsen: 00:38:43
Then the ambassadors would be responsible for their own little conference, and we would not be. The only thing that we would do for the ambassadors and say, “Here’s the general way that you can build a conference like this, here is our livestream, here is our logo, here is a website, we can help you with registration.” So that’s the only thing we really offer.
Margot Gerritsen: 00:39:04
That was, for us, a low cost, low effort way to scale up this conference. And the other nice thing of course is you’re giving your ambassadors a lot of ownership so they can build their own. And this just worked so well. So now we have over 500 ambassadors all around the world, and like you said in over 60 countries. And if we hadn’t had COVID, we probably would have had around 250 or 300 WiDS conferences around the world, and even with COVID happening we had our Stanford conference on March 2nd and two days later Stanford did not allow big conferences anymore. We were even thinking at the time, should we just cancel Stanford WiDS and in hindsight maybe that would have been better. I’ve always been worried about WiDS having actually been one of these conferences that helps spread COVID more. We weren’t as knowledgeable I think about COVID very early on in March as we are now.
Margot Gerritsen: 00:40:19
But anyway, we still did it this year but many of our conferences happened a bit later in the year and they either went to fully virtual conferences or they canceled and they will do them next year. But so we created in 2016 and ’17 this hybrid form of livestream, virtual conference, and then regional live events, and now a lot of also regional virtual events.
Margot Gerritsen: 00:40:48
What has been so amazing to me is to see this being picked up in countries that you may not expect, we’re really big in the Middle East, in Saudi Arabia we were, I think, the very first conference that women could go to unaccompanied because there were only women at the conference.
Kirill Eremenko: 00:41:06
Wow, very interesting.
Margot Gerritsen: 00:41:08
In Japan, this was the first time women really got together for a conference in this sort of computational area. We have them in Africa, we have them everywhere but Antarctica, so we should work on that. We should have a-
Kirill Eremenko: 00:41:25
That’s so lovely.
Margot Gerritsen: 00:41:27
Yeah. So it’s been a really surprising success, and we’re still a very small team, we are three co-directors and we have a really small budget, actually, to run this, but we are really making great use of this amazing work done by these hundreds of volunteers around the world. So it’s been so surprising to me.
Kirill Eremenko: 00:41:54
That’s awesome. When is the next one and how can somebody participate?
Margot Gerritsen: 00:41:59
So the next one will be International Women’s Day 2021, and then we are actually going to have a 24 hour event around the world, so it will be a 24 hour marathon where we start in the Asia-Pacific, so we’ll kick it off in New Zealand, this conference. Then we travel around the world, so we go to… so we have the Asia-Pacific and China and India, then we go to Europe and the Middle East, then we go to the Americas, and then we end up with live streaming from Hawaii.
Kirill Eremenko: 00:42:36
Hawaii.
Margot Gerritsen: 00:42:37
Yeah, so that’s a really, I think starting in New Zealand and ending in Hawaii is really fun. And then so we’re going through these different regions in the world, and we’ll have content delivered from Stanford for 24 hours, but that content will be supplemented by content in each of these different regions, from some of the regional events.
Kirill Eremenko: 00:43:02
Amazing.
Margot Gerritsen: 00:43:03
Yeah. It will be a really fun thing to do and like you said, this last year we had over 100,000 people participating in the WiDS conferences, and the nice thing also about it is because most of the conferences, the WiDS conferences, they are recorded, and so we have hundreds and hundreds of videos of women talking about their work right now. So it’s made a difference. But we don’t just do the conference, we also now have a podcast, that I’m hosting.
Kirill Eremenko: 00:43:37
What is it called?
Margot Gerritsen: 00:43:39
Well, we are incredibly creative so we call it the WiDS Podcast, so it’s really just called the Women in Data Science Podcast. We also have a datathon that we’re running, a global datathon for a couple of months before our WiDS conference hit Stanford, and then at this Stanford central conference we always announce the winners.
Margot Gerritsen: 00:44:05
So last year we had several hundred teams from around the world working on this datathon, and we have a bit of an outreach program now too to high school students. So we’re growing, we want to do more, but it’s still really the three of us running this so we try to grow pretty slowly.
Kirill Eremenko: 00:44:29
Yeah, wow, wow. That’s fantastic. What are the main messages you aim to spread with this conference?
Margot Gerritsen: 00:44:40
So we started because we really, really wanted to promote these fantastic women doing this really great work, but we always say we aim to support, inspire, and educate. So those are the three active action verbs, if you like, the action words for us. One of the big things we hope is that some of the barriers to participation by women, and we are really very inclusive, so when I say women, anybody really is welcome. Anybody who identifies as a woman, anybody of any gender, any background of course can participate in our datathon, can participate by going to our conferences [inaudible 00:45:31].
Margot Gerritsen: 00:45:32
But we’d really like to make people see that there is no reason for anyone not to enter this field, that you don’t have to fit a certain profile to be successful in this field. And the reason I’m emphasizing this, Kirill, is that I’ve seen over the three decades now or so that I’ve been in the field of computational sciences, there’s so many women or people from a different cultural background who somehow feel that because they do not fit the stereotype data scientist, or the stereotype computer scientist, or the stereotype mathematician, that they will not be successful and they don’t have what it takes.
Margot Gerritsen: 00:46:23
And that’s so sad because there is an enormous amount of talent out there and we’d like to break down those barriers for people from other countries, from people from other genders, from other racial backgrounds, and so that’s really my main motivation for doing this.
Margot Gerritsen: 00:46:40
Here’s one thing that I hear a lot. For years, teaching also in many places around the world, I’ve heard young female students say, “I’m just not good enough in mathematics or in computing or in statistics or in general technical field to really make it in this world, to be successful, however you define success, to be able to contribute. I’m just not good enough.”
Margot Gerritsen: 00:47:11
Then you ask them, “Why do you say this?” And they say, “Well, that’s how I feel, that’s what I’ve been told, I see boys or young men around me ahead of me, I feel they have this innate ability, I don’t think I have this innate ability.”
Margot Gerritsen: 00:47:35
So that’s one big myth that you have to have this really strong innate ability to be successful, and that’s not true. Innate ability helps, it’s certainly helpful, but it is not necessary. And a lot of people who believe in that, they feel that the first time they don’t do so well in the course means that they will just never be successful.
Margot Gerritsen: 00:48:04
But if you believe in a growth mindset and if you just stick with it for a bit, you often break through and you grow in this and you can reach a level of understanding that’s very high. But unfortunately if you have these fields where people say only with an innate ability can you be successful, and by the way math and computing and data science is not the only field, the same thing happens in finance and business to some extent, the same thing happens in philosophy, there is this notion that you need to have innate ability.
Margot Gerritsen: 00:48:38
And you combine this with the stereotype threat as Claude Steele, one of my colleagues at Stanford, names it, that men have this innate ability a lot more than women, then you’re creating barriers to entry for women that are just unnecessary. So that’s what I’m really hoping that this conference shows people, that there’s so many women that you can see there that are successful, young women and girls and also women that are in the field and still feel like a little bit of an imposter say, “Hey, there is absolutely no reason for me to fear this. There’s so many examples of women that are really good. I can be that way too.”
Margot Gerritsen: 00:49:26
Whether you come from a math background or stats background or come from the social sciences or the humanities or from the field of medicine or law or earth sciences or biology, it really doesn’t matter. You can become a great contributor to this field. And here’s the other reason why I’m so gung-ho on this. Data science, as you know very well, has become so critical. Many, many decisions that we’re making in the world right now, be it political, be it economical, be it technological, many decisions are data-based now. So it’s data-driven decision making.
Margot Gerritsen: 00:50:14
So that means that right now we’re setting the stage through data analysis for the future. And if you do not have a good representation of the population around the table making those decisions, doing these sort of analyses, you’re really missing out. That’s not good. So first of all, I would really like to see a much higher percentage of the people making those decision be women, I’d like to see much higher percentage come from different countries. It would be crazy to have companies in Silicon Valley really determine the trends of the future now in so many different ways.
Margot Gerritsen: 00:51:01
It is crazy to have a company like Google, which is founded on a certain culture, make decisions for our internet-based life, decisions that involve ethics and ideas of fairness with this western culture only, for the whole world. To me, it’s totally crazy. So this needs to be democratized in the sense that it needs to be globalized and we need people from all sort of backgrounds around this table. And I may sound a bit that way too but I’m a little frustrated that after decades of being in this field myself, so little has changed.
Kirill Eremenko: 00:51:52
Why is that? Why has so little changed?
Margot Gerritsen: 00:51:53
I think because of these persisting myths. When I was a young student in high school, I was one of the very few girls in my high school choosing math, choosing physics and chemistry. I was often one of two girls, maybe. And then I went to university and there was, again, one of two girls, and often the only girl in some of the classes.
Margot Gerritsen: 00:52:21
When I asked female friends of mine why are you not also mathematician or why are you not going into computer science, they would say, “Ah, I just don’t have what it takes.” That same barrier is there now. Still we see girls in middle school and high school say, “I’m just not good enough.” I think this is just persisting because we allow it to be. We still have the stereotype threat, there are still teachers who believe this actually, there are still so many role models for girls who say this.
Margot Gerritsen: 00:53:02
When my son was in elementary school I went as a guest to one of his little math classes to talk about pi, this amazing number pi. So he was talking to these eight or nine year olds about pi, and I already saw at that stage that there were many more boys paying attention than girls, and in the advanced math class a year later there’s only five girls out of 25. At the age of nine, there’s absolutely no reason for that. Girls are very good at mathematics and there’s absolutely no evidence that they’re not as good as the boys, but it was already starting at that time, this idea that boys are just better and faster.
Kirill Eremenko: 00:53:51
Do you think there’s a preference, like girls may be… not as many girls want to be in mathematics, or do you think it’s mostly about the preconceived myths that they have?
Margot Gerritsen: 00:54:07
You know, I think there’s… First of all, I have to be really humble here because I just talk from my own experiences and so obviously I’m probably a little biased in this, but this is what I think. I think that these two myths I talked about, there’s this feeling that society has as a whole that you must have this innate ability to be successful, the feeling that men have this more than women, is incredibly important. Then on top of this I think a lot of girls and young women are, and this may also be cultural, this may have nothing to do with genetics or really, really original female versus male traits, but we see a lot of girls and young women be very motivated to contribute to the social good, to do good.
Margot Gerritsen: 00:55:12
There is this also misconception, I think, that the STEM fields and data science and computer science are dry and removed from reality, maybe not so practical, and not used for the common good. Which of course is also totally false. But people think about when we go to high schools, we see this for example, a lot of the girls say, “Why would I want to work for Google? They’re only doing… they’re creating apps or they’re so far removed from what really benefits society.” So I think that also plays a role.
Kirill Eremenko: 00:55:55
There’s like a [inaudible 00:55:56]
Margot Gerritsen: 00:55:57
I think so. And it’s always hard because people argue and say, “Well, when you look at nursing, there are very few men doing that. Are you up in arms about the lack of men in nursing as much as you are up in arms about the lack of women in data?” I think it’s terrible. I think men would be fantastic nurses and I’d love to see more of them. That’s not my field so I’m not up on the barricades for men in nursing, but I think in general society benefits when people who are at the forefront of any movement, be it healthcare or data driven decision making have a very balanced demographics. I think that is to the benefit of society all along. It’s the same with first responders. Know that’s why it’s better to have more women in the fire service, more women in the police force, more women certainly with policymakers, for all of those reasons it’s good to be balanced. Whenever you have people making decisions for future generations.
Margot Gerritsen: 00:57:14
So I think in this it behooves us to really try to dispel these myths and only if we’ve done that and we still see females choosing those other fields much more can we probably say hey there is something innate to women or men that makes them go this way or another. Right now it may very well be yes, that women say, “I just don’t want to be in this field.” But is that because they don’t like the work or is that because they don’t like the male-dominated culture?
Kirill Eremenko: 00:57:57
Gotcha. It makes sense, thank you. And you’re clearly doing your part to help with this problem. You’re spreading the word on your podcast, on guest podcasts like this one, on your conference. What can individuals do, regardless of gender, regardless of age and background. Just somebody listening to this podcast, what would your call to action be to this person to help encourage more women into the space of data science?
Margot Gerritsen: 00:58:31
Well, if you are a woman listening and any of this you think oh yeah, I have a little bit of that too, please challenge your own misconceptions. I know this sounds very judgmental, but if you are a woman and you have these thoughts of I may simply not be good enough, I may not have what it takes, start challenging yourself a little bit and say, “What evidence do you actually have for this?” Start turning this around.
Margot Gerritsen: 00:59:07
Most of the time for women and others, it’s because they’ve found some setbacks or they’ve felt that they failed a couple of times, or they’re slow on the uptake, or they felt a little bit out of their depth. And it makes them feel uncomfortable and they see this as a sign of not belonging. But really all it is is that you’re on a learning curve, and when… this is the thing that I did for myself many years ago. When you start really loving that feeling, hey I’m feeling a bit overwhelmed, I’m feeling like I don’t quite yet have what it takes, I’m still learning. Don’t compare yourself to others who are a little bit further on this learning curve, but just think about how wonderful it is that you are learning and growing and embrace that and seek that out even more, and know and get the confidence that you will learn and you will get better. And this is so beautifully captured in this notion of a growth mindset, and at the very start of… before we started the podcast, you asked me what’s my favorite book.
Margot Gerritsen: 01:00:24
Here’s a book that I think is phenomenal for this, and that’s the book Mindset by Carol Dweck. She’s a colleague of mine at Stanford, her last name is D-W-E-C-K, and this book Mindset has changed the lives of many young women that I’ve known, and also young men because it’s not just the women who sometimes feel like an imposter. This discomfort with learning is in many people, there is this idea that when it comes to computing or math, you have to get it instantly or you will never get it.
Margot Gerritsen: 01:01:08
And that’s nonsense. As an instructor I see this. Even the people around you who seem to know it all, they have themselves also gone through this learning curve and they also make mistakes. It may be a little bit of a culture in this field to not really admit to that too much, but reading that book changes a lot of people’s perceptions. So I really highly recommend that.
Margot Gerritsen: 01:01:40
So challenge yourself and if you feel uncomfortable with not being on top of something yet, turn it around to say how great, I’m learning, I’m growing. And in itself that becomes a goal. I’m at the stage now in my career that I’m actually a little bit uncomfortable with being too comfortable. If I understand everything and everything comes easy to me, that doesn’t feel right because that means that I’m actually not learning all that much.
Kirill Eremenko: 01:02:12
Yeah, makes sense.
Margot Gerritsen: 01:02:14
So here is another thing I always tell students as well, that what’s an expert? You strive to be an expert on something, you strive to be successful on something, you strive to be a leader in a field. So you want to be an expert. What is that? An expert is somebody who’s made every possible mistake. So you can only become an expert if you make all those mistakes.
Kirill Eremenko: 01:02:38
Oh, that’s a good way of looking at it.
Margot Gerritsen: 01:02:40
So you’ve got to learn and fail and learn from that and fail again and keep going and climbing up that big steep learning curve. In itself, that’s the wonder of learning, and so hopefully… that’s what I’m hoping, people get out to say hey… And yes, for some people they’re a bit rusty in mathematics when they enter this field, they may have never programmed this much, and these are skills that you have to do a lot of. I sometimes do these bootcamps for people and I say, hey, math and computing is just like sports. You don’t go and say I want to run the marathon, oh let’s not train for a whole year and let’s just go and run a little bit, I should be able to do a marathon. It’s the same with math, you have to keep doing it, you have to maybe do it daily or a couple of times a week and have to keep training these muscles in your brain, so to say, create these connections in your brain and they need to be fresh and you need to constantly replenish that and grow that.
Margot Gerritsen: 01:03:50
And you get rusty, and this is another thing that happens with people. They haven’t maybe done math for a while and then they go take a course or they try to take a professional course in Python or something else, and it doesn’t come fast to them. Said no, you’re totally rusty, you need a WD-40 course, you need to de-rust yourself a little bit.
Margot Gerritsen: 01:04:13
What I always tell is just keep at it. It’s just like starting running again after not having run for a year, it takes a while to get back into this and you have to do your mobility exercises and your stretching, it’s the same with your brain. So just keep at it and keep feeding this and keep computing, keep trying to write code, and the more you do it the more natural it will become for you and the faster you will be. Compare it to sports.
Kirill Eremenko: 01:04:45
Yeah, makes sense, makes sense. And anything in life, like if you want to master your mind it’s not like you’re going to all of a sudden become the expert at meditation, you’ve got to meditate slowly little by little over the course of weeks and months and years before you reach a level of oh, calmness in the mind and you can control your thoughts.
Margot Gerritsen: 01:05:07
Yeah. And this is something I’ve tried and I’ve always totally failed at. Totally. And I’m really envious of people who can do this and steel their mind. I need to go on a bike ride or something to really steel my mind. I need to be out of breath to steel my mind.
Kirill Eremenko: 01:05:32
Everybody’s got their own. Thank you. For those of our listeners who are women or who identify as women, one way to or the main way is to question your beliefs and question where is this belief coming from, why am I… if that belief exists in their mind. What can men do? How can men support their female colleagues and encourage them to enter data science or to thrive in the space of data science?
Margot Gerritsen: 01:06:06
Yeah. That’s such a good question. I think for men, what I’ve seen with my male colleagues, what really helped them, of course there is this thing that you have to think about whether or not you have any biases in how you look at people. There’s a lot of them. Some of them are conscious, some of them are unconscious, so you have to challenge yourself a little bit with that. Look around you, see who you’ve been hiring, see who you’ve been promoting.
Margot Gerritsen: 01:06:35
It’s very common for people to hire like people. You tend to hire people a bit like yourself or the culture in your company. And so that perpetuates this thing if you are in a company that’s 85% or 90% men, it’s more likely that men are hired. So challenge yourself a bit in that, say hey I might just even subconsciously falling into that trap.
Margot Gerritsen: 01:07:00
But the other thing is, I think for men, is to really understand that it is very different for a woman or somebody from an underrepresented minority to be in a male-dominated or white or Asian-dominated, because that’s often the case in these fields, a group. And a lot of men that I meet don’t acknowledge that, they say, “Ah, we’re nice, we’re open, we’re welcoming. You don’t have to feel different.”
Margot Gerritsen: 01:07:39
So what I say to them is go to a Women in Data Science conference. I have seen amongst fantastically supportive colleagues at Stanford, a big change in their understanding of what it is like for women and therefore their dedication to supporting women after they attended a Women in Data Science conference, and for the first time in their lives, they were the odd one out.
Margot Gerritsen: 01:08:08
See, for a lot of men in this field, they have never, never been in a group where they were different. They never had to think about it. And I just remember one of my male colleagues coming to the first WiDS, and at the time we had 400 people at the conference, in this big auditorium, and there were 90 men, plus or minus one, I forgot now. So I ask them all to stand up.
Kirill Eremenko: 01:08:36
Plus or minus Joe.
Margot Gerritsen: 01:08:39
Ah, Joe couldn’t make it, that’s right. Joe wasn’t there. One of them said that he had to get up halfway through the conference to go to the bathroom, so he got up and he said for the first time ever I saw people look at me when I got up and left, walking through the crowd to go to the bathroom. Now, as a woman and when you’re at the conference, and I’ve often been in that situation, when you get up and you stand out literally, people focus on you. Not because you’re a model or you’re so good looking, that’s not it. It’s just simply because you’re different.
Kirill Eremenko: 01:09:32
Oh hey look, there’s a woman at this conference.
Margot Gerritsen: 01:09:34
There’s a woman, wow. Where would she come from? So now it was Women in Data Science, hey there is a man. And so this guy, this colleague of mine said this was the first time, and that was the moment where it shifted in him where he said, “Wow, I’ve never felt this different.” And it made him feel a little uncomfortable. This was just, I said to him, “You’ve just experienced this for eight hours. Most of my female colleagues my age, we’ve experienced this for 35 years and it still happens.”
Margot Gerritsen: 01:10:13
So that sort of awareness, when you’re a man, said throw yourself in situations like this, go to Women in Data Science or go to a meetup for women in machine learning for example, go to a meeting with Py ladies or R ladies, there are so many. There’s women in AI, there’s so many events now. We’re absolutely not the only one. There’s so many opportunities for men to experience this also and open their minds to it.
Margot Gerritsen: 01:10:44
Then I would say for those of you listening in who can hire and promote people, hire more women. Be open. Hire people that do not look like you. Don’t worry about it. Give them, if you had any doubts, give them the benefit of the doubt, try them out, and look at these hundreds and thousands of women who have been speaking at Women in Data Science conferences or Py ladies, like I said, all these other women who are amazing. And look to them as well, look a little bit beyond your normal field of view.
Kirill Eremenko: 01:11:29
Absolutely. And with your first story, I have also a story, which happened to be this year. And I completely agree with what you’re saying. So my girlfriend and I were in Buenos Aires, I think at the start of March, and that’s exactly when coronavirus started becoming serious and everybody started realizing okay, this is a big problem.
Kirill Eremenko: 01:11:57
And while we were there, the first couple of days it was fine, you could go to a restaurant or anywhere. But then the official government, they didn’t know what to do yet. So there was no instructions. But people started to catch on, like if they’re tourists they may have brought coronavirus, because Argentina was still okay. And we clearly look different to everybody there in Argentina, so we were walking down the street and everybody’s looking at us like whoa, stay away from them, they might have coronavirus.
Kirill Eremenko: 01:12:31
And when we would go into a coffee shop just for lunch or for some coffee, we would sit down and the owners would come and ask us all these questions like why are you here, what are you doing? Eventually we got kicked out just because of the color of our skin. Kicked out, nobody else asked anybody questions, because they all looked different. But we stood out.
Kirill Eremenko: 01:12:50
So when we would go in, we’d get kicked out. In that moment I realized oh, this is what racism feels reversed. I have no problems with that situation because it’s understandable, it’s coronavirus, people are afraid, people are worried, so I don’t have any hard feelings, but I can just imagine if you lived your whole life feeling that way, always looked differently at just because of the color of your skin it’s a terrible feeling and it took for me, like before I intellectually understood it, but it took that experience to understand it on an emotional level, it’s a whole different level of understanding.
Margot Gerritsen: 01:13:23
Yes. Yeah. And how, I mean we’re both white and so I have the advantage of being a Caucasian woman in this field, which makes it easier for me. I know this, much easier for me, than women or different genders, non-males of any kind. And particularly from other types of background. So my black female colleagues, my Hispanic colleagues, know that, my Muslim colleagues. For a lot of them, they have a double whammy or a triple whammy.
Margot Gerritsen: 01:14:10
So I need to be really humble too. Have I seen sexism? Absolutely. And I’ve experienced any level, from just the microaggressions as we call it, to being ignored, to not being promoted, to being scrutinized more than others, to being sexually harassed, everything on the spectrum I’ve experienced. And that’s not great. At the same time, though, because I’m Caucasian, it’s been easier for me.
Margot Gerritsen: 01:14:51
So yeah, this sort of challenge, to put yourself there, even though it is very limited because you knew also it was only going to be temporary, and it’s the same with the men, they come to Women in Data Science, they leave it, they’re back in their own comfortable culture. But imagine being in that situation over and over again. But the first thing we can all do is to try to at least understand a little, and I would never claim I understand because I’ve been a woman in STEM I understand what it is to be a black woman or man in the United States, for example. I would never claim that because I think that’s at a different level altogether again.
Margot Gerritsen: 01:15:35
I can leave my work and go home and be the normal. So I think it’s really, really important for people who are not different to understand that, and I find that when you then understand even a little bit, empathy is created and people tend to be more open and more aware. Because what is also really difficult for a lot of, let’s just focus on women here, women in this male-dominated environment, is that it’s always up to them to change. It’s up to them to point this out, it’s up to them to be part of diversity committees and so on.
Margot Gerritsen: 01:16:15
And what we would really like to see is that we have men taking responsibility, ownership for this. So here’s another call to your listener. If you are in the company, for example, that wants to work on diversity, don’t ask the women to do that for you. Give that ownership to the men, send them to something like WiDS, and to give them feeling for what it is.
Margot Gerritsen: 01:16:55
Sure, ask advice from women, understand the literature and things, but take ownership. We have, in some cases, men for example now also stepping up and saying, “I am not going to be part of any manels. If I’m asked to be on a panel at a conference and everybody on that panel is a man, I refuse. I refuse to speak as a man at a conference with only male panel speakers.” These things really make a difference.
Margot Gerritsen: 01:17:27
And I hear so often people say, “Ah but you can’t do this, that’s reverse discrimination, that is too far the other way. Why would we invite women just because they’re women.” And they said, “We just have to hire the best, that’s all. And if that happens to be all men, what’s wrong with that?” And then I say, “Challenge yourself a little bit on how you define the best. If by the best you also mean the person you are most comfortable with, who looks most like you, who has the same sort of definition of best, who is maybe as aggressive or has the same sort of way of speaking or communicating, then you need to really challenge how you valuate best here.” And that is what often happens.
Margot Gerritsen: 01:18:21
Then if you want to hire or promote or get somebody to speak at your conference and you just can’t find any, reach out to us. We’ve got binders of women, right, as we said in the States. Anyway, that’s a little political joke, but we’ve got many, many, many lists and videos of amazing women, so we can help you with that.
Kirill Eremenko: 01:18:49
Awesome, awesome. Thank you, Margot. This is very, very valuable and unfortunately we’re going to have to wrap up here because we are short on time, but I greatly appreciate the insights and the offer to help those, people find the right speakers or right women. I definitely will be approaching you from our conference, DataScienceGO, to ask for some speakers. I think it’s a great noble cause that you are pursuing here. So thank you very much.
Margot Gerritsen: 01:19:20
Thanks, Kirill, it was really fun to chat with you.
Kirill Eremenko: 01:19:24
Awesome. Before you go, where are the best places for our listeners to contact you or find out more about WiDS and follow your work?
Margot Gerritsen: 01:19:32
Ah, so you can send me an email, I’m super easy to find. To find me, all you have to do is google Margot with a T, M-A-R-G-O-T, Stanford. And they will find me, they’ll find my email address. They can always email me and I will respond. I’m also on Twitter and Facebook a little bit but I’m not as active as I maybe should be on those sort of platforms, but email is the easiest.
Kirill Eremenko: 01:20:04
Okay. Is it okay to connect on LinkedIn as well?
Margot Gerritsen: 01:20:07
Absolutely. Yeah. Absolutely. Anybody who wants to connect just send me an invite, send me a little message with it as well. I tend to respond to LinkedIn mostly if people have a personal message and a reason to connect with me. But if you say you listened to Kirill’s podcast show, I will connect with you.
Kirill Eremenko: 01:20:30
Awesome, awesome. And WiDS, what’s the website for WiDS?
Margot Gerritsen: 01:20:37
widsconference.org, W-I-D-S conference, one word, dot org. And again, they can just google Women in Data Science Stanford and I’m sure they will find it [inaudible 01:20:48].
Kirill Eremenko: 01:20:48
Amazing. And the closest one you said is, what, March 2021?
Margot Gerritsen: 01:20:52
March 2021. International Women’s Day, March 8th, 2021, it’s going to be the WiDS at Stanford and the 24 hour WiDS event.
Kirill Eremenko: 01:21:02
Amazing, amazing. Well, again Margot, thank you very much. It’s been a huge pleasure speaking with you today.
Margot Gerritsen: 01:21:08
Yeah, thanks very much Kirill.
Kirill Eremenko: 01:21:15
Thank you very much everybody for tuning in today, hope you enjoyed this podcast with Margot and got some valuable insights. My favorite part, I had two favorite parts, my first favorite part was about principal component analysis, what a beautiful explanation Margot provided. It was extremely powerful and valuable and also very just clear. And my second favorite part was that notion of being on the other side of racism or sexism or being in the minority, how it feels is different to how we think it feels. So if you have never experienced that feeling, then as Margot recommended, try to put yourself in an environment where you are the minority, where you will feel on an emotional level what it’s like to be different, to be the odd one out.
Kirill Eremenko: 01:22:13
So try that out for yourself and perhaps that will help you even more to inspire others and help them thrive in this space, regardless of their background, regardless of their gender, and any other factors that really don’t and shouldn’t matter.
Kirill Eremenko: 01:22:31
So there you go, that was our podcast with Margot, and as usual you can find the show notes at www.superdatascience.com/407, that’s www.superdatascience.com/407. There you’ll find a URL to Margot’s LinkedIn, URL for Women in Data Science conference, and any materials we mentioned on the show as well as a transcript for the episode.
Kirill Eremenko: 01:22:53
And one final call to action, if you know any woman or anybody who identifies as a woman or as a matter of fact, anybody who could benefit from this episode, to help make data science more diverse, more inclusive, and more empowering to everybody, then send them this episode, send them, iIt’s very easy to share, send them the link, www.superdatascience.com/407.
Kirill Eremenko: 01:23:18
So spread the love and let’s make data science an amazing place to be. And thank you again to Margot for joining us for this podcast, and thank you for tuning in today. I look forward to seeing you next time, and until then, happy analyzing.
