Deborah Berebichez: [00:00:00] It's reminding ourselves that our measure of success is not what people think and how many likes we get, but how we measure our learning and our growth, just learning to appreciate how much you have grown and come forward is an incredible skill that you need to nurture and practice. And that is going to take you places, because the more comfortable you feel with failure and getting up again after small failures, the more success you'll have. As I always have said, the people who end up succeeding in life are not the ones for whom things come easy. They are the ones for whom obstacles are just something to transcend and the ones that get up every time that they experience a failure in their lives. And they keep going.

Harpreet Sahota: [00:01:06] What's up, everyone? Welcome to another episode of the Artists of Data Science. Be sure to follow the show on Instagram @theArtistsofDataScience and on Twitter @ArtistsOfData. I'll be sharing awesome tips and wisdom on Data science, as well as clips from the show. Join the free open Mastermind selection by going to Bitly.com/artistsofdatascience. Where I'll keep you updated on biweekly open office hours. I'll be hosting for the community. I'm your host Harpreet Sahota. Let's ride this beat out into another awesome episode. And don't forget to subscribe. Rate. And review the show.

Harpreet Sahota: [00:01:56] Our guest today is a physicist, data scientist and TV host through sheer force of passion, persistence and perseverance.

Harpreet Sahota: [00:02:03] She became the first Mexican woman to graduate with a physics PhD from Stanford University, where she studied under a Nobel laureate, Steven Chu. And she accomplished this despite growing up in a conservative community that discouraged young women from studying or pursuing careers in science by leveraging her expertise in scientific research and advanced analysis. She has helped many organizations automate the decision making process and uncover patterns in large amounts of data. And she's cultivated a specialty in drawing connections between the approaches used data science and the challenges faced by organizations. She's the chief data scientist at Metis where she leads the creation and growth of exceptional Data science training opportunities such as bootcamps, corporate training, professional development, as well as live online programs, all while also developing a world class Data science instructional team. She's also the co-host of Discovery Channel's Outrageous Acts of Science TV show and co-host of the TV show Humanly Impossible, produced by the National Geographic Channel. She's taking her STEM research initiatives globally, empowering and encouraging women in places like India, Israel, Mexico and Costa Rica. And her work has been recognized by the likes of the Wall Street Journal, Oprah, Dr. Oz, CNN, Ted, DLD and Wired magazine. So please help me in welcoming our guest today. A woman who's passion is to inspire young women and minorities to pursue careers in STEM. To think deeply, to be bold and to help others. Dr. Deborah Berebichez. Dr. Berebichez thank you so much for taking time out of your schedule to be here today. I really, really appreciate you being on the show.

Deborah Berebichez: [00:03:38] Well, thank you so much. That was a wonderful introduction and it's an honor for me to be here.

Harpreet Sahota: [00:03:44] Thank you. Thank you. So. So let's talk about your path into Data science. What sparked your interest? Where did you start and how did you get to where you are today?

Deborah Berebichez: [00:03:54] I just so I think it was a serendipitous path in that I didn't really expect to become a data scientist.

Deborah Berebichez: [00:04:02] I had never heard about the term. And maybe about 15 years ago when I had finished my PhD, I, I started working in Wall Street.

Deborah Berebichez: [00:04:14] Like many physicists, because I wanted to be able to get a green card and stay in the U.S.. And as you know, there was a strong connection between the financial markets and the PhD programs in physics and math and statistics across the country. And so it was kind of not even raised eyebrows. There were over a few thousand physicists working Wall Street. And so I finished two post-doctoral fellowships after Stanford at Columbia University and at NYU, at the Grant Institute in Applied Math and Applied Physics.

Deborah Berebichez: [00:04:55] And then I started working in physics and I realized that academia was a bit too isolating for me. And I wanted to communicate more with the public and evangelize different products and have an impact with my coding and what I was doing. I did computational physics, by the way. And so it was pretty close to Data science. I just would not we would just not call it that. But I had never realized that what I was doing was a very narrow form of Data science, meaning I was, you know, quite proficient with that particular aspect of machine learning.

Deborah Berebichez: [00:05:39] But when it came to Data science, which was much more vast than what I was doing. And so I was humbled by an experience I had at STRATA, the big data science conference when I was interviewed on video. And I think I said something that I, I, I've regretted saying ever since, which was oh, but come on, Data science is nothing new. You know, we have physicists and Wall Street people doing it for the past 50 years and nothing has changed. And, you know, I was proven wrong quite quickly because we definitely were analyzing things with different algorithms and we were analyzing different kinds of data that we never analyzed before, such as audio and text and images and what not. And so there there were a lot of differences. And also in Data science, you we required to translate the insights that were gained into quite, you know, lay and entertaining terms so that the stakeholders in a company could actually enact policies in a. And change things in vier the company into a different direction to gain success based on those insights. So that's how I started. I finished my appose socks. I worked in Wall Street for six years and then I realized that what I was doing in Wall Street was research and and again, do working with Data. It was the stock market Data to be specific. But I also knew that I wanted to have more of an impact in the world and do good for people.

Deborah Berebichez: [00:07:28] And at the time, I had been following my friend Hilary Mason, who's a renowned Data a scientist, and I loved her work.

Deborah Berebichez: [00:07:37] And I saw Cathy O'Neil and other people do use the Data science analysis that they did for bringing more ethics in into the world and and more visibility into under served communities when it came to doing data science work.

Deborah Berebichez: [00:07:59] And so I ended up wanting to connect education with Data science. And that's how I came about Metis, which is where I'm the currently the chief data scientist at. And we're a Data science training company where I've had the chance to not only train people by teaching a machine, learning bootcamp and create curriculum, but also where I've had the chance to do Metis for good projects like helping create a live map of needed and things during an earthquake that happened in Mexico about four years ago. And people could go to the map and in real time see what kinds of items or people were needed in different locations. So it's been a wonderful world of work where I can actually not only help people, but also educate companies and others in Data literacy. And that's what I loved about my work and Data.

Harpreet Sahota: [00:09:07] What's up, artists? Be sure to join the free open mass, my snack community by going to Bitly.com/artistsofdatascience. It's a great environment for us to talk all things Data science, to learn together, to grow together. And I'll also keep you updated on the open biweekly office I'll be hosting for our community. Check out the show on Instagram @artistisofsdatascience. Follow us on Twitter at @ArtistsOfData. Look forward to seeing you all there.

Harpreet Sahota: [00:09:36] That's pretty awesome. You've been able to apply your expertise in a number of different domains. I'm curious as to your thoughts. What do you think is going to be the next big thing in Data science in, say, the next two to five years?

Deborah Berebichez: [00:09:51] Yeah. So I think that a lot of Data science has been slowly partitioning into more and more specific professions.

Deborah Berebichez: [00:10:04] We have tried to capture a vast amount of things under the umbrella of Data science, and that has not worked well because companies have been hiring Data scientists, some of whom have expertise in Data management.

Deborah Berebichez: [00:10:21] Others more in sophisticated algorithms like deep learning and others more in a more statistical base. Data analysis. And so I think people want to know what they can get out of data science.

Deborah Berebichez: [00:10:38] And so we're seeing the proliferation of dashboards and easy platforms like Tableau that are going to be able to be used within an organization with very little training.

Deborah Berebichez: [00:10:53] That is pretty much anyone will be able to have access having an initial training to the data that a company has. And people will have insights at every level. So we're going to see that.

Deborah Berebichez: [00:11:05] And those people are going to be translators or bridging bridges between the executive levels of the company and the companies Data.

Deborah Berebichez: [00:11:15] And so we will need to hire very kind of heavy engineering background or Data science backgrounds for those things. At the same time as algorithms, sophisticated and more complex algorithms become successful at solving certain problems, we're going to see more people hire specific bands within Data science.

Deborah Berebichez: [00:11:39] That is somebody who is exclusively an expert in NLP algorithms or in visualization techniques and whatnot. And so I think that more and more jobs are going to open.

Deborah Berebichez: [00:11:54] But we're going to they're going to require more specific skills and more training in certain areas, as well as people from less technical backgrounds having access to more commodity sized platforms.

Harpreet Sahota: [00:12:12] So this vision of the future you have. What do you think is in the separate the great data scientists from the good ones?

Deborah Berebichez: [00:12:19] Oh, that's a good question. I think somebody who has the skills that I call critical thinking will definitely advance. Way more than the good data scientist. So I think we could define a good data scientist as somebody who is able to efficiently manipulate, clean and gain insights from a data set that have actionable metrics that can propel a a company or an institutions business forward, whereas a great data scientist will be somebody who can think outside the box and outside of the established algorithm in both. Go back to the basics and make sure that the statistics are correct, which a lot of people don't think about now and not be deceived by, say, the sample that gather the data. The agenda behind the data source out of the company that's providing the data and whatnot and really gain deeper insights by creating an algorithm that specifically tests what they know they want to test with with the metrics that are as specific as to the errors that I get propagated with statistically measuring only a sample of the population.

Deborah Berebichez: [00:13:43] And we're not really paying attention to how at every step of a data science project, we can unintentionally or sometimes intentionally propagate these errors and misuse data science to gain insights. That are eliminating from our goal the version of a comprehensive truth, so to speak, like we can, you know, test voter Data set, by eliminating unconsciously the opinions of certain minorities or certain other political views. And, of course, then gain insights that are not actually representative of what the political ecosystem is.

Harpreet Sahota: [00:14:30] So what do you think would be kind of the scariest application or the scariest abuse or misuse of data science machine learning? Let's say in the next two to five years?

Deborah Berebichez: [00:14:42] Yeah.

Deborah Berebichez: [00:14:42] So I think a lot of data science is now being used for analyzing sensor data.

Deborah Berebichez: [00:14:49] And by that, I mean that the data that we analyzed no longer lives in a screen in, a computer, but it's actually living in sensors that health companies have that people are going to wear on their bodies, for example, to measure the amount of medicine they have to take with a sort of a sticker that punctures the skin and they measure the amounts of a particular medication. The blood, those are out there already. We're going to measure we're going to put sensors in offices, in manufacturing plants, in clothing, in our scales, in our bathrooms and closets. You know, go figure in like all kinds of things. And when all of these Data, which is many times communicated to other platforms on security, gets used by different companies with different agendas. That's where we're going to see a cross section of Data being abused and being misused. So, for example, while it may be useful for parents across the world to have cameras in their homes to check their babies at night if they're moving or crying or what not. We see that there is a website for specifically for Internet of Things products called Shodan that is out there where you can find lots of data from open cameras that have been put not only in baby rooms, but in back of restaurants and whatnot that are just simply open because the security protocols have not been put in place. And these Data have, of course, the potential to be abused. So, you know, we're going to see that. I think that in the Internet of Things and in sensor Data, we're going to have to be very thoughtful about how we engineer our future and the communication between all these platforms.

Harpreet Sahota: [00:16:46] Yeah, that does sound really scary, especially about the baby monitors that my wife and I just had her first baby three weeks ago. Oh, yes. Thank you. Thank you, I have a baby monitor in his room. So right after this, I could check the security settings on that thing. OK. So on the flip side of it, now, let's talk about, you know, in what ways do you think Data science will have the biggest positive impact on society in the next two to five years?

Deborah Berebichez: [00:17:11] Well, I do think that a Data science done well is going to expand the reach of pretty much everything that companies do. So, for example, banks tend to lend money to a particular set of people. And in the past, that set of people was that had access was typically the one that was sort of set, had a social profile that was pretty recognizable, had a certain salary and lived in a certain area and behaved in a specific way. Whereas now we are gaining more insights into how minorities or people with different profiles behave when borrowing money. And so more companies are going to feel more comfortable, for example, lending to people who don't fit exactly that profile.

Deborah Berebichez: [00:18:03] We're also going to have more transparency. So I have a friend, for example, Dan Ariely. He's a behavioral economist and very cleverly, he's looking at data of how companies treat their employers. And probably we're going to be able to create a financial index of all the companies that treat their employees better.

Deborah Berebichez: [00:18:26] May may well be the ones that are profiting the most. And so we're going to be able to invest in them more. And so we're going to shine light in two different areas of society that where invisible before a project that I really like, for example, was MIT's looking at trash and what happens to trash. So this is an IoT project where they basically put G.P.S. chips into three thousand trash objects in Seattle and they followed them for two to three months. And they saw that in the beginning, the trash objects went everywhere. Shoes and cartridges and bad or old batteries and whatnot.

Deborah Berebichez: [00:19:09] But after a few months, they saw very clearly how some markets developed and some print cartridges, old ones, get reassembled and go to Mexico because it turns out that it's expensive to buy new ones. And so they inject them with ink and they reuse them.

Deborah Berebichez: [00:19:27] So there's a market for that. Whereas alkaline batteries into Africa for a similar reason, they get recharged and they get resold in different market. And so by just adding a chip, a sensor and making that Data visible, we can inform governments, trade unions and other stakeholders about how to manage these markets, better reduce tariffs or open ways for these markets to thrive and so on.

Deborah Berebichez: [00:19:57] So I think the most amazing things that are going to happen are giving transparency to industries and to communities of people that otherwise in the past have remained quite invisible.

Harpreet Sahota: [00:20:10] That's really interesting. You definitely have given me and will definitely our listeners a lot to think about and lot to research. Thank you for that. So I'm trying to think of a clever way to sneak this question in, but I think the best way to do it is just ask. In doing my research, I came across. Came across. You mentioned in your obsession with Tycho Brahe. Can we talk about first of all can you give us just like a.

Harpreet Sahota: [00:20:34] A synopsis as to who Tycho Brahe is or was and what your obsession is with him. And the biggest thing that we all can learn from him.

Deborah Berebichez: [00:20:43] Yes. Thank you for asking this question. In fact, you are going to find this funny, but I have an eleven month old son whose name is Teacup because I am so in love with the figure of Tycho Brahe that I named my son after this famed astronomer, Danish astronomer.

Deborah Berebichez: [00:21:01] So it's a very unique name, I think. There were when we checked the census, there were only five kids in the US with that name. So hopefully it'll be a trend in the future. But anyways, yes.

Deborah Berebichez: [00:21:15] So when I was growing up in Mexico City, I was a very curious and inquisitive child. And I always wanted to know more about the universe and math and physics. But I was discouraged from pursuing a career in STEM and science because I was a woman. And in this conservative community, they told me that I shouldn't really dedicate myself to something so male dominated and that I should pick something easier and more feminine, like marketing and other things.

Deborah Berebichez: [00:21:43] And so I started to read behind everyone's back books about obscure physicists like Tycho Brahe, who did amazing things in their lives, but did not count with a very healthy social life because I just thought that would be me in the future. So Tycho Brahe was this Danish astronomer.

Deborah Berebichez: [00:22:02] He was a noble man and actually visited the island, which is now in Sweden and was part of Sweden. Nowadays, the island where he lived, I visited his beautiful observatory underground then. It was just an incredible moment for me. But basically he was he didn't have such a great personality. There were rumors that he fought in a duel and the king expelled him from Denmark. And then he after fighting in this duel, he lost his real nose and he had to wear a nose, an artificial nose made for him out of copper and all that. However, as a scientist, he was magnificent. He out of eight thousand observations of the sky with a naked eye. Remember, we did not have big telescopes back then. He had this. He built the most precise equipment. At the time, there were instruments to be able to track the movement of stars and across the sky. And with you know, they say big data is everything. Well, with a very small data set, only 8000 observations. He was able to pretty much give them the foundations for deriving the laws of planetary motion that Kepler later did with Tycho's Data.

Deborah Berebichez: [00:23:16] In fact, it is said that Kepler kind of stole the data and used it to make his assertions and his axioms. So, you know, Tycho was a very interesting figure, extremely dedicated to his craft of being an instrumentalist and a scientist. And so I always grew up thinking that maybe I'll be like him, I'll be rejected by society in some way, but I'll have my observations and I'll have my science with me. And that's that's how I felt isolated from the world for liking Data and scientific discovery. But somehow in love with the process. And he became a kind of a silent hero for me. And now I think I've taken Tycho Brahe's teachings to the next level because I didn't I wasn't content with keeping them secret. I actually became a science communicator because I have very strongly seen how much people are held by increasing their their science literacy.

Deborah Berebichez: [00:24:19] And I am a very strong supporter of making people learn and educate others in the basics of science so that we can become empowered citizens and know more about the world.

Harpreet Sahota: [00:24:32] A very beautiful story. Thank you so much for sharing that with us. 

Harpreet Sahota: [00:24:38] We talked about the wish he'd make a great Data scientist. And you mentioned critical thinking. Can you talk us about what does it mean to you to be a critical thinker?

Deborah Berebichez: [00:24:50] Yes, critical thinking to me is about questioning authority, things.

Deborah Berebichez: [00:24:56] I teach a course the art of deception in Data science. And I can't tell you how many examples I take from publications that are as reputable as The New York Times. And then they put a graph, they say, you know, doubling of income in this area or doubling of the number of voters. But when you look at the actual absolute number, it went from like or a doubling of efficacy of a medicine. And it goes from like .001% to .002% percent. OK. Yes, you doubled it. But in objective measures, it's is still not very efficacious at all. And, you know, yes you doubled the income. But it went from, you know, one dollar to two dollars. What does that say? So there are there are ways of presenting Data, of visualizing it that is meant to intimidate or to deceive people or sometimes it's done even unconsciously. And so the y axis, you know, is cut. At a certain point, them they don't display the full size of the bars or some more. It's present that something's presented. Not in a logarithmic scale when logarithmic scale would be more appropriate, etc..

Deborah Berebichez: [00:26:11] And so for me, critical thinking starts there. OK, let's. Let's look at what the news are giving us. Who is benefiting from this kind of insight and this kind of result? What's in their agenda? Who owns the data? Who's doing the analysis? Like, is this medical equipment company paying for the data analysis that showed that that particular equipment is is, you know, super efficacious for for, you know, there are health study or not. So you're asking questions about the Data, asking questions. Did this take into account communities of underserved communities to conclude this is this take into account women's voices and whatnot, like all kinds of questions that you can ask. And we at Metis have a course on Data science literacy that equips people with the skills and the tools to basically analyze what they see and all the visualizations that we see everyday on LinkedIn, on Facebook, on marketing campaigns, and and allows us to to gain the proficiency in being able to discard lies from the truth. And that, to me, is critical thinking, or at least where it's starts.

Harpreet Sahota: [00:27:33] Do you have like a actionable tip that our listeners can take with them so that they could cultivate a habit or mindset of critical thinking for themselves?

Deborah Berebichez: [00:27:42] Yeah, absolutely. I always say go back to the sources every time you read something, pretend that the person who wrote it or the agency that put it out there has absolutely no authority to not trust authority. And that's something that Feynman a physicist used to say all the time, do not trust authority. And not only that, but Feynman also said it's very easy to fool other people, but it's actually easier to fool ourselves. So make sure that you recognized the biases that you have about the world and what you want to be truth so that you don't blind yourself to the actual results of a Data analysis. And so when people do Data or even when people read about things, make sure that the evidence presented is something that makes sense to you. Like check the sources, check, check.

Deborah Berebichez: [00:28:38] You know why people are forming this conclusions and you'll become a much more informed reader. At first it's going to be difficult. Like you're just going to want to trust the sources. But after a while, you realize how many of the things we get bombarded with are actually not true, or simply the research has been done incorrectly. And just equip yourself with a basic statistics course and learn what are the typical ways we get deceived. And we do statistics the wrong way so that you can be armed and be the voice that critiques that out there as, hey, this Facebook post here doesn't seem to be true. You know, the research does not support these conclusions actually because of this and this and that.

Deborah Berebichez: [00:29:25] And I can't tell you how incredible and empowering it is when you can actually become that voice of reason.

Harpreet Sahota: [00:29:33] So why are humans so bad at appreciating or conceptualizing probabilities?

Deborah Berebichez: [00:29:39] I think because we live in a world that is extremely complex and we have so many factors that could be credited for bringing about some result. So even in science, you know, when I see something fall or something behave in a certain way in the natural world, it's a multifactor problem because it could be caused by angular momentum here, but it could also be. And, you know, the way we've dealt with that in my TV show, because during the filming, we would actually have to discern what were the most important factors governing some crazy thing that someone posted a video about on YouTube like they did. They roll down a mountain by being inside a truck tire and not injure themselves because of their rate of rotation was so fast that it overcame the tires inclination to to fall sideways or, you know, like all kinds of things. And so when we see the world and so many factors that could be causing something, it's very difficult for humans to entertain those equations because they're very hard and long and difficult to solve. So I think that's why humans, human brains, although we have good intuition, we also make a lot of mistakes in guessing where people where things are coming from.

Harpreet Sahota: [00:31:09] So in statistics, it's kind of like the adage with like enough data anything can become statistically significant. So for Data scientists out there who are working with vast quantities of data, why is it important that we cultivate this intuition for what probability represents?

Deborah Berebichez: [00:31:25] Because you can be very easily fooled. There is a story that Maurice Samiah, who I admire very much. When he was still working at Google, there is a story that she was very meticulous and she tested exact shade of blue in which the the Google search engine would report the links. You know, when you search something, every link has a beautiful. Blue color and she tested so many shades of blue. Like maybe hundreds to figure out through AB testing which one was the best one. You know, the one that that had the best response from people. And she felt. And the the reports in the press said, OK. So they found the perfect shade of blue. However, there's something in statistics. Too many repetitions can actually lead you to come up with a false accuracy because you test so many shades of blue.

Deborah Berebichez: [00:32:23] That by pure probability, you can end up getting very high response for one particular shade. And it's not because it's true, but it's by fluke. You had you know that in that run, that effect happening.

Deborah Berebichez: [00:32:38] So we're you know, if you question the results, we're really never sure if that was actually the best shade of blue or that particularly that particular run happened to give those results.

Deborah Berebichez: [00:32:53] So it's very important to question that in a business and the question that decisions, not what people are doing with Data. So make sure that you understand the statistics and what's happening when you test something with an algorithm that it's very, very useful so that you can actually report the right conclusions to your boss or to management at the executive suite and not report your bias or what you think should be happen.

Harpreet Sahota: [00:33:27] Are you an aspiring Data scientist struggling to break into the field? Well then, Check-Out DSDJ.co/artists to reserve your spot for a free informational webinar on how you can break into the field. That's going to be filled with amazing tips that are specifically designed to help you land your first job. Check it out. DSDJ.co/artists.

Harpreet Sahota: [00:33:53] Thank you for that. Shifting gears a little bit here. I'm wondering how you view Data science. Do you view it as an art or science?

Deborah Berebichez: [00:34:00] That's an excellent question. I think I caution people who view it as an art because although there is definitely an art to it in the skill of, you know, doing exploratory data analysis and whatnot, I do think that the best data scientists are the ones who exploit the sort of scientific process of going through, meticulously proposing or systemically proposing.

Deborah Berebichez: [00:34:28] Algorithm after algorithm. Each one more complex, but trying many different ones in a systematic way to figure out which one can read the data best.

Deborah Berebichez: [00:34:38] Especially when we have when we get to more obscure algorithms like deep learning.

Deborah Berebichez: [00:34:44] I like to think that people who are taking that approach have already exhausted all the more open and linear and clear algorithms that exist out there because many, many times we can like in physics. The best solution is the more elegant and economical one in the sense of, you know, the one requires the least amount of data and effort. And so I think Data science is is quite scientific and in that sense. But definitely because it's not getting a solution to certain problems, there is an art to it, which means that Data in the end is subjective. We capture it by subjective means, like polls and questionnaires and what not. And different Data a prize of a stock is influenced by a lot of human subjective measures and behaviors. And so no matter how we try to make something fit into a particular system, it may not because it doesn't follow rules like the natural world.

Deborah Berebichez: [00:35:55] And so we have to accept that there is an art to it. And to the extent that people are open to multiple interpretations. I think we as long as we know the limitations of our results in Data science, then we are able to call it a science with huge error bars.

Deborah Berebichez: [00:36:15] Sometimes. 

Harpreet Sahota: [00:36:16] Thank you. I really, really enjoy that response. I'm curious how how do you think the creative process tends to manifest itself in Data science?

Deborah Berebichez: [00:36:26] You know, I write. I like to write. And so when I write, which I consider creative process, I first write my thoughts in very concrete, in a concrete list of items, so to speak. From there, I try to connect those thoughts.

Deborah Berebichez: [00:36:42] And after that, I try to create a story that links those thoughts together and puts them in a context so that people can relate to the story better.

Deborah Berebichez: [00:36:55] And that's an art, because I could come up with, you know, various ways of connecting those thoughts and various ways of putting a context to them. And so that's a very artistic process. So in the same way, I think a good data scientist sits down, looks at the data and perform some exploratory analysis in the beginning and says, OK, you know, what can I do with this data? What what are the basics? What? What? For example, let me do a Linear analysis just very quickly and see what it says or let me see. You know, let me just visualize the data. Does it cluster, you know, amongst a certain direction? What is it telling me? And so when we do that, we're able to see more of that artistic process take place because it goes from very basic facts.

Deborah Berebichez: [00:37:46] And looking at the data to creating a story that can encompass what's happening with the data and what the data is telling us and how it's going to effect change when we use the insights that came from it.

Harpreet Sahota: [00:38:00] For people out there who are trying to break into data science and maybe they feel like they don't belong or they don't know enough or they aren't smart enough. Do you have any words of encouragement for them?

Deborah Berebichez: [00:38:12] Yes, I think Data science has really brought the barriers down for for doing things that are technical in the world.

Deborah Berebichez: [00:38:24] You know, it used to be that. You really needed a very powerful computer to do astrophysics and you needed to be, you know, extremely good at manipulating multiple dimensions in your head to be able to visualize things. And now with Data size, I think anybody can get into the field. We have high school girls that I've mentored that know how to analyze Data extremely well.

Deborah Berebichez: [00:38:52] You know, they may use simpler tools like Excel or something else.

Deborah Berebichez: [00:38:57] But but they're able to gain insights from Data very, very quickly. So know that there are various ways in which you can become a data scientist from a book to learning on your own learning.

Deborah Berebichez: [00:39:12] In university, we at Metis, have amazing bootcamps where we also find jobs for people. And don't be intimidated by it. You're probably going to be better at some aspects than others. But the field is vast. And, you know, some companies need Data science interpreters who are better at communicating the technical details. Other companies need people who are better at engineering and others at the algorithms and what not. And so there's always room for different backgrounds, different skills, and it's all about gaining confidence from where you came from to where you can be. Not about comparing yourself to the the skills and successes of others.

Harpreet Sahota: [00:39:54] So in those moments where we feel like we're failing or failures, we want to give up because it's hard upskilling and learning so much to get into Data science. What can we do to feel like a hero?

Deborah Berebichez: [00:40:07] You're probably referring to a TEDx talk again, how to be a hero when you feel like a failure. And I think it's reminding ourselves that our measure of success is not what people think and how many likes we get, but how we measure our learning and our growth based on what our goal was before we took on an enterprise.

Deborah Berebichez: [00:40:34] Just learning to appreciate how much you have grown and come forward is is an incredible skill that you need to nurture and practice. And that is gonna take your places, because the more comfortable you feel with failure and working and getting up again after small failures, the more success you'll have. As I always have said, the people who end up succeeding in life are not the ones for whom things come easily. They are the ones for for whom obstacles are just something to transcend and the ones that get up every time that they experience a failure in their lives and they keep going. Those are the ones that get to the end.

Harpreet Sahota: [00:41:23] Absolutely love it. Yeah, I was referring to that TED talk. You've got some of the most inspiring TED talks that I have listened, I listened to all of them. They're really, really amazing. I'll be sure to link them to our listeners as well. And that point about the obstacle is very important. I've been on a stoicism tip recently, reading a lot of Seneca and Marcus Aurelius, things like that.

Deborah Berebichez: [00:41:46] Ah my husband loves the stoics.

Harpreet Sahota: [00:41:48] That's good. They're amazing. I feel like I've been a stoic for quite some time, but I never put a label to it. But I think what Marcus Aurelius says is the impediment to action advances action. What stands in the way becomes the way. So the obstacle is the way. Do you have any tips for those who are coming from a non-technical background? Maybe they've been in their career for, you know, maybe 10, 20 years in a I.T. type of role. But they're coming to Data science and they're coming face to face with these foundational concepts for the first time and are feeling a bit intimidated.

Deborah Berebichez: [00:42:27] Yeah. I mean, I can tell you, one of the most successful Data scientists of all time, my friend Hilary Mason, had hired an English major and other non-technical people to be Data scientists at her amazing company, Fastforward Labs.

Deborah Berebichez: [00:42:45] So I always say that, you know, Any, it doesn't matter what your background is, you can have the most incredible insights and you can be very, very good at working with Data. Another thing is, in our bootcamps, we sometimes get applicants that have non technical college majors or we even had an applicant who join our boot camp in San Francisco. He was 17 or 18 years old and he didn't even have a university degree and he actually wanted to do the bootcamp instead. And to our surprise, he is the first one in his cohort to get a job as a data scientist. And so the advice is to not, again, not compare yourself to others and believe in yourselves because everybody brings a unique gift and a unique perspective to a company. And you may have what's called domain expertise, which is incredibly useful. You may just be an amazing collector off, I don't know, old cars or old paintings or something that happens to be what the company is trying to do. And so you may come with an Data science skills are not quite developed yet, but you may have some domain expertise that happens to be incredibly useful.

Harpreet Sahota: [00:44:06] What would you say would be like the biggest myth that people tend to hold in their heads about breaking into Data science? And would you mind debunking that for us?

Deborah Berebichez: [00:44:14] Yeah, I think one of the biggest myths that I've seen is that people doing data science are geniuses and are like really good at every aspect of Data science.

Deborah Berebichez: [00:44:28] And they're have a very skilled portfolio of doing all kinds of Data science. One that's actually not true when you really look at very good data scientists. Yes, some of them are are quite good at many different aspects of Data science.

Deborah Berebichez: [00:44:45] But believe it or not, the vast majority of people working in Data science are quite good at one, two or three areas and types of algorithms. You know, you you shouldn't strive to acquire very deep skills in every area. You should strive to know how to speak about all the various aspects and possible algorithms within Data science for sure. Have a general enough knowledge, but do not strive to become an expert before you get your first job. Just, you know, or become proficient at, one, two or three areas out of the hundred that exist. And, you know, the ones that are your favorite, the ones that lend themselves well to your background or your skills or what you want to do. And market yourself a lot to get your your first job. And in that job, you're going to grow and you're going to learn more in-depth things about your specialty, but also more things in general about Data science.

Harpreet Sahota: [00:45:49] Absolutely. Hundred, ten percent agree with that advice. The same advice I tend to give my mentees as well. I think now would be an excellent time for us to, you know, talk about Rupesh Story. So would you mind. Would you mind sharing that with share with our audience?

Deborah Berebichez: [00:46:05] Sure. So I said I grew up in a in a conservative community that discouraged me from doing science. And so when it came time to pick a topic, I picked philosophy in Mexico because I was told it's philosophy. Ask questions about the world.

Deborah Berebichez: [00:46:21] Kind of like physics. And then. I started studying philosophy for two years, and my hunger to do physics and learn math did not go away. So behind everyone's back, I applied to schools in the U.S. because I learned that you can have a double major. And I won a scholarship to Brandeis University, which was very helpful because my parents couldn't have afforded to be an American school at the time. And I had only two years because I was a transfer student to pursue my degree. And when I got to Brandeis, I realized that I had the courage to take an astronomy class, which was a very generic course. And there I when I finally had the courage to study physics.

Deborah Berebichez: [00:47:05] I had the fortune to meet the graduate student who was a teaching assistant from this class. His name is Rupesh and he was from India and he was studying a PhD in astronomy. And he and I became good friends because he said I wasn't the typical student that just wanted a good grade and the homework. You said my eyes would open up and and would be bright asking with questions about the universe, and the planets and whatnot. He was the first person to truly believed in me. So when I told him one day that I didn't want to die without trying to do physics, he helped me meet with the graduate student committee. The head of the graduate student committee and who also was the head of the physics department and said, you know what? Somebody else did this before you. Many years ago, Ed Witten, who's, by the way, the father of string theory and physics, and he said if by the end of the summer and the next two months, you're able to master this material. And he handed me a book which was a vector calculus in three dimensions, which was an alien language to me at the time. Of course, I didn't even remember algebra. Well, and he said, if you're able to master this book and we'll test you on it, we'll let you skip through the first two years of the physics major so that you can cram in the next two years and finish with a BA in physics and so Rupesh was incredible because he decided to devote his entire summer to mentor me.

Deborah Berebichez: [00:48:36] And he tutored me every day. We didn't really have a lot of time because it was two months to cram. The first two years of math, physics and calculus and algebra and everything, classical mechanics and thermodynamics and all of that. And at the end, it was a successful thing because I was able to graduate with highest honors from physics and reason why tell the story of Rupesh is because I always wanted to pay him for all that he did for me.

Deborah Berebichez: [00:49:03] And he cited a story of when he was growing up in India, in Darjeeling, an old man who was climbing up the mountain where Rupesh lived to teach him and his sisters math, English and the tabla, a musical instrument. And whenever Rupesh's family wanted to pay this old man, the old man said no. The only way you could ever pay me back is if you do this with someone else in the world. And basically, that's what Rupesh did with me. He said I could only pay him back by making this my life's mission, which is to inspire and encourage other people and other minorities who, like myself, feel attracted to STEM, but who for some reason being social or financial, feel that they cannot achieve their dreams.

Harpreet Sahota: [00:49:51] I feel like you've repaid Rupesh hundredfold with all the work that you've done.

Deborah Berebichez: [00:49:56] Thank you. You should let him know.

Harpreet Sahota: [00:49:59] So, you know, a lot of data scientists, they're working on projects and they might feel some hesitation or some fear because they're trying to make the project perfect before releasing it to the world before they get on. Do you have any tips for anyone who is in that mindset?

Deborah Berebichez: [00:50:15] Yes. If you ship it when it's perfect, it's too late. That's what they say about startups. You need an MVP, basically mean viable product. So, you know, usually if you depending on the project, of course, if you're creating a product that's going to be used by consumers, it does need to have a certain level of ability and perfection to it. But if you're analyzing data, I would say, you know, at optimal 20 percent of the ways already are way ahead of where you started.

Deborah Berebichez: [00:50:47] So even with an exploratory data analysis can discover so many things that I say, communicate early and perfect later, always communicate with your team, communicate, communicate your initial results and, you know, get a bunch of people to think about what they mean and what are the next steps, because instead of working like crazy on perfecting something and you're going down the wrong path, it's best to catch those potential misleading paths from the beginning and just, you know, get to experiment and see what the Data is telling you from the get go so that you can have even less perfect results. But at least once that are true and represents reality.

Harpreet Sahota: [00:51:32] To switch gears here again, I was wondering if we can speak about your experience being a woman in tech and if you have any advice or words of encouragement for the women in our audience who are breaking into or are currently in tech.

Deborah Berebichez: [00:51:47] So unfortunately, we haven't advanced as much as we've wanted to in the area of equity and equality for women in the tech world and in other areas.

Deborah Berebichez: [00:52:01] And so my advice is somebody once told me off, if I act as if meaning.

Deborah Berebichez: [00:52:09] Don't let the perception and the stereotypes that have formed your unfortunate biases govern what you do and how you behave. Act as if you're confident. Act as if even if you don't feel confident yet and things will happen for you and things will be there. And also, I want to say that there are different modes of leadership.

Deborah Berebichez: [00:52:37] It doesn't mean that women have to imitate the typically stereotypical male modes of leadership and be assertive in that way.

Deborah Berebichez: [00:52:50] What it means is that you can be, you know, still yourself with your own qualities and maybe you're a quiet leader. Maybe you'd like leading people by writing and not by leaning in in a meeting and making lots of comments in meetings. There's always a way for you to become a leader. Seek that way and seek mentors out there that believe in you and they're going to propel you to the top.

Deborah Berebichez: [00:53:16] Because the ones that get to the top, probably need in this time somebody that helps them, that believes in them, that helps them, you know, with a few steps in that corporate ladder here, there.

Harpreet Sahota: [00:53:30] What could the Data community and men in the Data community do to foster inclusion of women in Data science and AI?

Deborah Berebichez: [00:53:39] Make them role models, promote women, seek women to hire them for the roles that you have, put technical women in highly visible roles. If you have a conference that that your company is going to sponsor or you are going to attend, send your technical woman to speak. Make women role models visible. That's the best you can do.

Harpreet Sahota: [00:54:05] And you mentioned this a little bit earlier, but I was wondering if you could talk to us a little bit more about the initiative. You led to develop the first high school curriculum for Data Science for girls.

Deborah Berebichez: [00:54:15] So Moody's Analytics contacted us and we were a group of seven women that wrote and executed the first Data science curriculum for high school girls of underserved backgrounds. And it's now being deployed by Girls Inc, which is a wonderful organization that helps to equip young women, especially if underserved backgrounds, with the skills to succeed in life and in their professional lives. And so we teach in high schools. We train teachers to we train them to do our specific curriculum in Data science. Of course, with tools that are freely available. And then we find internships for these high school girls and they get to work in various companies doing some simple Data analysis or visualizations or even working with Excel. That's not free, but large companies have it.

Deborah Berebichez: [00:55:12] And then we we get them to experience what it's like to be part of the ecosystem in tech and data science. And that definitely changes the life, their lives, because by having that experience and that example, there are way more likely to select a technical career in the future.

Deborah Berebichez: [00:55:34] And, of course, that improves their professional lives in many ways.

Harpreet Sahota: [00:55:39] I think it's so cool that you started that up. Last question before the lightning round here. What's the one thing you want people to learn from your story?

Deborah Berebichez: [00:55:47] You can make your dreams come true no matter what. Do not believe in what people think of you and what your abilities are. To always seek for that inner voice that tells you what you want to do and believe in yourself. Because we're all prone to fall prey of what the media tells us and what even our well-meaning family and friends tell us. But only we know what we can achieve because we are the ones that are going to put all the effort into making our dreams come true.

Harpreet Sahota: [00:56:24] So let's go ahead, jump into lightning round here. What is your Data science superpower?

Deborah Berebichez: [00:56:30] Oh, being so detail oriented, I love, like checking code and I can people who miss a comma or something, I'll find the error it like faster than anyone.

Harpreet Sahota: [00:56:45] So what's the most outrageous act of science that you've come across.

Deborah Berebichez: [00:56:50] Oh my God. Our entire show is very it just exercises full of them.

Deborah Berebichez: [00:56:54] Like, you know, I've seen people in my show do outrageous and insane things that I never thought a human could do. Like jumping, you know, flying under the, close to the surface of a mountain with a with a flying suit and nothing more.

Deborah Berebichez: [00:57:13] Not a parachute. Nothing like that. Or people jumping off of buildings and, you know, all kinds of things.

Deborah Berebichez: [00:57:21] So I think it just takes an hour to look at an episode of our show and then you'll realize that nature can be outrageous.

Harpreet Sahota: [00:57:30] So let me get a little little deeper philosophical here with the next couple of questions. OK. What would you say is the most fundamental truth of physics that all human beings should understand?

Deborah Berebichez: [00:57:41] Physics is not about facts that the way they teach us science in school. Many times about, OK, choose the multiple choice. And there's only one right answer. It's not the right way to do science. At the same time, of course, it's not. The science has multiple answers, but that physics and other scientists are an iteration of ever increasing approximations to the truth, meaning that you know where people Newtonian mechanics is still true. Even though Einstein modified it for certain context in which we are travelling at the speed of light.

Deborah Berebichez: [00:58:20] But it doesn't mean that classical mechanics as elucidated by by Newton, is not correct in our sort of everyday world. So the same way when people get frustrated by scientists and they say, oh, you know, who can believe scientists these days? One day they say COVID gets, transmitted by touching surfaces, and then a month later, they say, no, it doesn't get transmitted by surfaces as much.

Deborah Berebichez: [00:58:49] So how can we believe that? Well, you know, what happens is that's the process of science. Science is a discovery.

Deborah Berebichez: [00:58:56] And it constantly changes because we're constantly finding new things that present a a more comprehensive and a more colorful, colorful picture of reality. So if physics changes is because we're going forward, it's because we are finding new ways of applying our laws to the world and we're finding cases in which they don't operate that way. And so science is not about facts.

Deborah Berebichez: [00:59:24] It's about discovery and an ever increasing, more comprehensive view of reality, nature.

Harpreet Sahota: [00:59:34] So what would you say is the most mysterious aspect of our universe?

Deborah Berebichez: [00:59:38] Wow. There are so many hidden things in physics like dark matter. Like why is the universe expanding, like all this string? Is there some validity to string theory, to the string being composed of.

Deborah Berebichez: [00:59:53] I'm sorry, the universe being composed of little strings and the model that we have, the standard model of all the particles that exist, as you know, is.

Deborah Berebichez: [01:00:06] Is there something else that we're missing out there that could. Because of all the forces that we know of the world. Gravity is incredibly weak and it's just so weak compared to the nuclear forces and the strong and the weak force. And when that is just a mystery to know why gravity does not appear, that's at the smaller scale. So how can we unify all these forces to have a more unified view of the universe by incorporating the force of gravity? I think that's a huge mystery.

Harpreet Sahota: [01:00:44] What's an academic topic outside of Data science that you think every Data scientists should spend some time researching on. 

Deborah Berebichez: [01:00:52] Critical thinking.

Harpreet Sahota: [01:00:53] What's the number one book, fiction or nonfiction that you would recommend our audience read and your most impactful take away from it?

Deborah Berebichez: [01:01:00] Richard Feynman's, what do you care what other people think?

Harpreet Sahota: [01:01:05] I love it. Feynman is freaking amazing. I love that guy. So if we could somehow get a magical telephone that allowed you to contact 20 year old Debby, what would you tell her first? First can tell us, you know, 20 year old. Where were you at? It will give us a little bit of context and then what would you say to her?

Deborah Berebichez: [01:01:26] I was in Mexico City studying philosophy because a lot of people had told me that I wouldn't be able to do physics and I wouldn't be able to do quantitative things.

Deborah Berebichez: [01:01:36] So I would tell that Debbie to believe in herself, to trust that there her desire desires to to do physics were valid and and came from a valid place of curiosity. And what's more important in life is to to pursue those dreams that we have, even if we are not the best of them. I'd rather do that than stay doing something that comes easy to me that I'm comfortable at satisfying me because my true desire was to learn or do something else. That's what I would tell her.

Harpreet Sahota: [01:02:17] What's the best advice you've ever received?

Deborah Berebichez: [01:02:19] My husband is a physicist too, and his adviser told him something that has taken me years to understand. What he said is hold your water. And I never quite understood what he meant by that. And my husband always says it's like if you are in a conversation where you you are adversarial or you coming too up to a point where it's going to just end up frustrating both parties and you're not getting kind of moving forward in the conversation, just maybe sometimes it's better to listen and to stay quiet and hold your water. That is, don't engage. Don't try to be the one who's right. Don't try to be the one who wins an argument. But hold your water. Like, stay back and and try to understand the other person better.

Harpreet Sahota: [01:03:12] So what motivates you?

Deborah Berebichez: [01:03:14] What motivates me is my children is is creating a world where my three and a half year old told me the other day, Mommy, can I do coding after the math?

Deborah Berebichez: [01:03:27] And I almost cried because I'm teaching her coding with code emoji, which is a very sweet app or website, and they teach coding to very young kids through emojis.

Deborah Berebichez: [01:03:40] And I don't know, it just brings me hope that they can inherit a world where the work of all of the marvelous women before me and after me are making it possible that we can succeed and we can occupy many more positions than the ones that were available to us in the past.

Harpreet Sahota: [01:04:05] Do you have any tips for me as a brand new dad? Who's a non-stop skeptic raising a child in a family of believers.

Deborah Berebichez: [01:04:12] Oh, yes. So I would recommend buying books by Ruth Spiro on science like quantum mechanics for babies, their so cute, and thermodynamics for babies, and basically educate your children to ask questions from the get go.

Deborah Berebichez: [01:04:33] Don't ever tell them that because you are their parent. They should believe in you. They should. I know. I know. It's much harder to bring them up that way. But tell them to question you. That does not mean you going to give them permission or you're gonna agree with them. But they still have the right to question the why. And I tell you, the teachers of my daughter have told me, even at her young age, that what they admire the most about my daughter at the young age of three is that she is able to reason through anything. If they tell her not to do something in school, she responds very well. As long as you explain to her, why not? No, because this is dangerous or because it's somebody else's turn or whatever way.

Deborah Berebichez: [01:05:22] But reasoning is a capacity that they can develop from an early age. So try to build that. I have talks with her where she says, Mommy, we need to have a talk and we try to reason through things to arrive at a conclusion. And I may still disagree with her and, you know, put the foot on the ground and say, that's your all you won't watch videos tonight, but at least she gets a chance to reason with me of why she thinks she should be watching videos.

Harpreet Sahota: [01:05:49] That's so awesome, I love it. What's what's the song that you have on repeat right now?

Deborah Berebichez: [01:05:54] David Bowie changes. I love that song.

Harpreet Sahota: [01:05:57] Awesome. How could people. How could people connect with you? Where can they find you?

Deborah Berebichez: [01:06:02] Everywhere in social media. My Twitter and Instagram handles are Debbie Bere. That's d e d d i e and then B as a boy. E. R. E. Debbie Bere. You can also find me on Facebook and LinkedIn and hopefully I'm Tick-Tock in the near future.

Harpreet Sahota: [01:06:22] Dr. Berebichez thank you so much for taking time at your schedule to be here today. I really, really appreciate you coming on the show. Thank you.

Deborah Berebichez: [01:06:28] Thank you so much. What a fun classic interview I really, really enjoyed.