SDS 227: Enhancing Your Mobile Gaming Experience With Data Science

Podcast Guest: Sarah Nooravi

January 17, 2019

What game on your phone have you considered the best and has never let you down?

Today, we talk to Sarah Nooravi from MobilityWare to find out what role does data science play in enhancing your mobile gaming experience. Aside from this, she also shares her exciting career journey – from culinary to engineering to data science – so start tuning in!
ABOUT SARAH NOORAVI
Sarah Nooravi is the Marketing Analyst of MobilityWare, a company that develops games for iOS and Android. Sarah is also ranked 5th in the list LinkedIn Top Voices 2018: Data Science & Analytics. She also has a background in Mathematics, Economics, and Engineering.
OVERVIEW
Welcome to the SuperDataScience Podcast!
I’m pretty sure that today’s episode will get your full attention and definitely give you a bunch of inspiration. I’ll be joined by Sarah whom I’ve met on DataScienceGO. Sarah has lots of story in store for us including, but not limited to, her distinctive career background, her current work where they use data science in marketing mobile games, and diversity in data science.
“Question everything. Answer with data.” This is what is written on Sarah’s LinkedIn bio and what she believes in. She proves in the first part of the discussion that data can answer all questions. Sarah invites everyone to ask the logical questions and find answers to them through the use of data. If something does not make sense, make sense of it by rationalizing it.
On the next part of the discussion, we chat – almost like a heart-to-heart conversation – about her career background. The huge jumps from the career paths that Sarah has chosen every now and then were quite peculiar. She dreamed of excelling in the world of culinary, and, later, she studied mathematics, economics, and engineering. After those paths, she locked her eyes on data science and had, ever since, been acing it by being valuable to her company, the gaming industry, and the data science community.
She managed to learn data visualization and machine learning. While she was learning, she was also organizing meetups and hosting webinars. She shared her knowledge and taught new skills to anyone in the data science community.
Sarah is the sole designated marketing analyst for a gaming company. She shares how she works with the entire team from scoping to implementation to productization. Discover also what Sarah’s role is in enhancing your mobile gaming experience with data science.
Lastly, we give light to the importance of diversity in data science. It’s proven that diversity is what mainly drives a successful company. Through diversity, we get the best ideas and the best solutions. It’s best if a company is a representative of what they expect their customers are.
IN THIS EPISODE YOU WILL LEARN:
  • “Question everything; answer with data.” [06:01] 
  • From culinary to mathematics to mechanical engineering to data science. [11:48] 
  • Learning data visualization and machine learning. [25:48] 
  • Hosting meetups. [27:50] 
  • The responsibilities of a marketing analyst. [32:35] 
  • Data Science and Mobile Gaming. [38:00] 
  • Implementation and Productization. [42:08] 
  • Diversity in Data Science. [46:58] 
ITEMS MENTIONED IN THIS PODCAST:
FOLLOW SARAH
EPISODE TRANSCRIPT

Podcast Transcript

Kirill Eremenko: This is episode number 227 with Data Science Influencer Sarah Nooravi.

Kirill Eremenko: Welcome to the SuperDataScience Podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur and each week we bring 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: Welcome back to the SuperDataScience Podcast, ladies and gentlemen, and today I’ve got a very exciting and fun and positive episode prepared for you. I just got off the phone with Sarah Nooravi and I definitely don’t think that I’ve had this many laughs on an episode of this podcast before. It was lots of fun and be prepared for a very, very energetic and positive episode.
Kirill Eremenko: What you need to know about Sarah is that she’s a Data Science Influencer with tens of thousands of followers on Linkedin and Sarah inspires the Data Science community through her articles, webinars, mentorship meetups and many other ways that Sarah engages in the community. She inspires data scientists to constantly learn and grow in their careers.
Kirill Eremenko: In this podcast, we talked about three main things. First of all, Sarah’s background and how she got into the space of Data Science in the first place. Be prepared for some very peculiar detours here starting from the world of culinary and becoming a chef and going all the way to to the world of nuclear fusion. Then after that, we talked about a specific case study or a specific use case of Data Science in Sarah’s current role and you’ll find out how Data Science can and is used for marketing of mobile applications. Very interesting case study and I’m very excited for you guys to check it out and find out, get a glimpse into this world.
Kirill Eremenko: Finally in the third part of this podcast we talked about diversity in Data Science and what we as a community can do to help inspire everybody regardless of their gender or ethnicity to be successful data scientists. There we go, we’ve got a very exciting podcast coming up ahead. Can’t wait for you to check it out and without further ado, I bring to you Data Science influencer Sarah Nooravi.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast, ladies and gentlemen. Super excited to have you on the show today and we’ve got a very special guest joining us from Irvine, California, Sarah Nooravi. Sarah, how are you going today?
Sarah Nooravi: Very good, Kirill. Thank you so much for having me. I’m excited.
Kirill Eremenko: I’m super excited and it was really cool meeting you at DataScienceGO. We were just chatting about this before, how we were, I think the first time we bumped into each other when we were putting those stickers under the chairs, completely not expecting, I wasn’t even expecting to do that, but yeah. Thanks a lot for helping out, I think it was a fun night we had, with all those stickers under the chairs to facilitate the conference, it was really appreciated.
Sarah Nooravi: Oh, it was a lot of fun.
Kirill Eremenko: Yeah. In a brief recap, I know we chatted about this just now but just for the sake of our audience, what do you think of DataScienceGO?
Sarah Nooravi: I think it was a very, very, it was a very well put together event. I think my initial thoughts were like, “Wow,” like the map that we had in the very, very front when you come, it’s a super impactful moment when you realize that you’re bringing people from all around the world to come to this event to meet each other, to network, to be a part of something really big. I think overall, I want to give it to you for putting together such a great event.
Kirill Eremenko: Thank you.
Sarah Nooravi: For bringing of the energy, for basically getting people excited to get into a space that is not that easy, right? It’s not that easy to break into it and so having that supportive community that’s going to help you whether it’s through networking and jobs or whether it’s through resources or support or mentorship, having that community to lean on is super important. Props to you for bringing people from all around the world and creating such a successful event.
Kirill Eremenko: Thank you, thank you and definitely right back at you because I couldn’t have done it without you guys. Like we had quite a few influencers there and as you pointed out correctly just before the podcast that we really leveraged this community that already exists on Linkedin on data scientists and thank you all so much for your shout outs and Tarry, also Eric, Randy, Favio, your shout outs on Linkedin to get everybody excited about the event. I think in overall it was really cool, and the diversity, right? That is a part that we’re very proud of, that we had an abnormal for this industry percentage of women or minorities represent at the conference. I think that is also to do with the committee that you guys have built up as influencers in this space. Once again, thank you for supporting the event and making it all possible and making it all happen.
Sarah Nooravi: Of course, of course, thank you.
Kirill Eremenko: All right. Well today we’re talking about your journey in the space of Data Science and your career and what you’ve done. I want to start off with something I noticed on your Linkedin which is really cool and it sounds to me like it’s your personal motto, “Question everything, answer with data.” That is such a powerful statement. How did that come to be?
Sarah Nooravi: Yeah, actually so I was thinking about it because Eric had a good one. Shoot, I forget what his was now.
Kirill Eremenko: Eric Weber, right?
Sarah Nooravi: Yeah, Eric Weber had a good one and he kept getting called out for it. I was like, “Oh, okay, I need to come up with a good one too.”
Kirill Eremenko: I think his is, “I learn everyday.”
Sarah Nooravi: “I learn everyday,” yes, that’s what it is. I was like, “Okay, so what is it that I do on a daily basis,” right? He learns everyday and I was like, “Wow, that resonates with me too.”
Kirill Eremenko: Just copy it. You should have just copied it.
Sarah Nooravi: Yeah, I was like, “Should I just take it?” I was like, “What do I do?” I was kind of thinking about it and I was like, “You know, I question everything. Everything has to come down to a logical question and answer. Okay, but why are you doing this? Why is that happening? Let’s get to the root of the problem,” right? At the end of the day, especially in businesses and even in personal relationships, it comes down to, “Okay, well historically what has been going on? How can we answer this with data?”
Sarah Nooravi: I feel like it falls right into my personality in my day to day and what I love to do is just be inquisitive, be curious and then don’t let people’s gut or their instincts lead what strategy ends up happening or what decisions get made but let that be based on something tangible. Data is tangible, actions are tangible. Yeah, I think it fell right into place and I like it.
Kirill Eremenko: Gotcha, gotcha. That’s a very apt way of putting it and totally agree, you got to use data to answer all those questions. Sometimes though, interestingly, I was speaking to Vitaly, my mentor, and sometimes he says that even as a consultant he sometimes uses, he relies on his heart as a separate entity for answering questions. Sometimes, you can call it gut feel, you can call it like following your heart, but sometimes even if the data doesn’t align with what his heart is saying, sometimes he’ll follow his heart. What are your thoughts on that? That’s a bit of a controversial comment there.
Sarah Nooravi: Yeah, that is, especially when you deal with stakeholders who want the data to only mimic what they want, what their heart is telling them or what their gut is telling them. They only feel good or it’s like a reassurance of like, “Oh, well when the data matches what my gut is saying then, okay, I’m good but when it doesn’t then I’m going to basically argue with you until it matches what I want.” It’s a little controversial. I think that especially on the, let’s say from the analysts’ side you have to have a hypothesis of what you think the data’s going to tell you, right? Because that’s how you’re going to approach the problem.
Sarah Nooravi: From the stakeholders’ standpoint, they’re going to question everything you do and everything you say until it kind of aligns with what they want it to say, which is good depending on who you’re working with. It just depends on the scenario, but that one’s a hard one, I think.
Kirill Eremenko: Yeah, yeah.
Sarah Nooravi: Hopefully data is the most objective way that you answer any question, right, so you would hope that if the data is vetted and you know where it’s coming from, it’s cleaned properly, the way it’s being collected is vetted and then your approach is sound then really, you should be trusting the data. Or at least you can modify a little bit of what your gut is telling you to align.
Sarah Nooravi: You know, it’s funny though. We as humans can convince ourselves of anything, right? Have you heard of this where the data could be, you could come up with research or data that tells you one story and maybe initially you don’t agree with it but then you can rationalize it. “Oh, okay, yes, it’s saying this because of X, Y, Z,” but then later discover something wrong with the data and it tells you, you come to a 180 degree different conclusion. Then you’re like, “Oh, okay, but I also see how that can [inaudible 00:10:39].”
Kirill Eremenko: Yeah, yeah, yeah. I totally know what you mean.
Sarah Nooravi: It’s very interesting how us as humans, we can take what the data’s telling us and come up with a story as to why it is this way or the other way.
Kirill Eremenko: Yeah. I’ve had that in my life. Kind of similar to the placebo effect when you’re given medicine and you’re told that it will help you with your high blood pressure or whatever else and in reality it’s actually not real medicine. It’s just an empty capsule but your brain creates a story for itself and convinces itself on a physiological level even to lower the blood pressure and what not. Interesting, interesting.
Kirill Eremenko: Okay, well Sarah, tell us for the benefit of our listeners who don’t know you yet, which is probably, I would say there’s a lot of our listeners who do know you. You’re a major influencer in the space of Data Science with tens of thousands of followers on Linkedin, but for those of our listeners who haven’t met you yet, can you give us a quick overview? What is it that you do and how did you get into the space of Data Science?
Sarah Nooravi: Sure. I never know how to answer this question. I like to start from the very beginning which is maybe too far back.
Kirill Eremenko: When you were born.
Sarah Nooravi: Because it’s interesting. You talk to people and their journeys into how they ended up where, especially into this field of Data Sciences, so vastly different. Mine started actually without even a desire to be in anything technical. I actually really, really aspired to be a chef growing up.
Kirill Eremenko: Really? Wow.
Sarah Nooravi: Yeah, I really wanted to go to culinary school.
Kirill Eremenko: You couldn’t be further away from being a chef by being in Data Science.
Sarah Nooravi: It was a really big passion of mine at the time when I was younger. Once I realized that that was not going to be the direction I would go I really fell into my love of mathematics and just logic in general.
Kirill Eremenko: Hold on, hold on, you just skipped a whole, I don’t know, massive part of your life story. When did you realize that it’s not the path you’re going to go down?
Sarah Nooravi: Okay, do you want to know the truth?
Kirill Eremenko: Absolutely, always, of course.
Sarah Nooravi: Because there were a few colleges around me that offered culinary programs and the one, when I realized what the curriculum had, and this is actually very interesting when you start talking about your passion for anything in life, are you willing to do the dirty stuff before you get to the most exciting stuff? The first classes that they wanted me to take were about sanitation. I was like, “What?” I was like, “No, I just want to start learning how to cook cool things and the creative side and the artistic side and the different flavors and this and that.” That part of it just completely turned me off and I was like, “No.” I guess I didn’t have anyone around me that was going to push me in that direction anyway. It was going to be 100% my own motivation into it and I fell off the cliff right there.
Kirill Eremenko: Wow, that’s crazy. Actually I heard that about chefs. I read an article once, I think it was about Jiro Ono who is the top sushi chef in Japan. Basically he or whoever this article was about, one of the top chefs there about sushi, when they went to learn to do sushi and they have this master who is teaching them how to do it, they weren’t actually allowed to touch the rice for, I kid you not, for 20 years he was not allowed to actually touch the rice. He had to watch, clean the place, do, feel and sense everything and now he’s the best chef in Japan with dozens of restaurants and super highly rated.
Sarah Nooravi: That’s when you think about whether someone really, truly wants to do that, pursue that career or pursue a certain hobby. You have to really enjoy every aspect of that job or of that hobby. Even the practicing, even just sitting around and watching other people do it, learning from other people’s techniques, doing every aspect of that career or of that job. Yeah, for me that was the point that I was like, “Okay, moving [inaudible 00:15:22].”
Kirill Eremenko: I love it, I love it. You just had to look at the curriculum to realize, “Nope, moving on.”
Sarah Nooravi: Then at that point, then I started college and I was like, “What’s the common thread of what I enjoy doing?” It really and honestly when I think about how I got into really enjoying math, it was through an English class. In English they teach you how to logically put together an argument. It’s very structured and it’s a logical flow of ideas. Through that logic, A then B then C, I realized that it’s really this underlying logic that was a passion for me and then I found it through mathematics. Then I studied math, econ, I minored in statistics. I mean I’m jumping ahead. I took a detour into mechanical engineering thinking I was going to go into the renewable, into the energy sector.
Kirill Eremenko: Oh yeah, as you do, just a casual detour into mechanical engineering. Wow. This is interesting. All right, well what made you take the detour into mechanical engineering?
Sarah Nooravi: When I graduated, I really only saw myself pursuing one job and it was really odd that I stuck to this one particular job that I wanted upon graduation, which was for a company called J-PAL. I was very excited about their mission. I was excited about what I would be doing with them. At that point, Data Science wasn’t really hyped up at that point. Maybe a little bit but it was barely trending upward and so I was looking for more of like a statistician’s job, designing experiments and helping the world in general. I wanted to have an impact. When I realized that that job was on the East Coast, so geographical limitations, I was like, “I’m not going to move to the East Coast and deal with snow.”
Kirill Eremenko: Makes sense. You were always in California, correct?
Sarah Nooravi: California, yeah. Then I decided, “Well, what’s the next best thing?” Because I moved back home and something about me, I always enjoyed teaching and tutoring. I took on a tutoring job. I took on a tutoring job, I moved back home, took on a tutoring job. I was getting paid almost nothing and I was like, “What am I doing with my life? What is the next best move for me?” I found through a class that I happened to take that I really enjoyed thermodynamics and I loved physics and I loved, like maybe my way of contributing would be through something like nuclear fusion just completely blew my mind.
Kirill Eremenko: Wow, so you went from being a chef to nuclear fusion. You’re a person of extremes, aren’t you?
Sarah Nooravi: I mean when you get excited about something it really, it’s that type of excitement that you can have. Like, “Oh my God, I want to have an impact and I want it to be in this,” right?
Kirill Eremenko: Yeah, yeah. I totally agree. I get excited about nuclear fusion every morning. I wake up and I’m like, “Nuclear fusion today, yeah, tomorrow laser physics.” I completely get your point. It’s just like the topics you pick are so out of the blue. Very interesting. Keep going, I’m having so much fun. This is really cool.
Sarah Nooravi: I mean I’m being totally transparent and honest right now.
Kirill Eremenko: Thank you.
Sarah Nooravi: I was amazed at the idea of creating a mini sun in your home, right. Like that was the future of nuclear fusion and I was like, “Well, how do you make that a possibility?” I started applying to master’s programs in mechanical, well there’s a whole story about how I ended up finally deciding on mechanical that is just hilarious but then I wanted to marry it with public policy. Just because I realized that in engineering, I didn’t end up actually setting public policy but I feel like someone who has those two skillsets can actually make a difference because you’ll realize that the way budgets get split for different research projects especially in government have to do a lot with understanding public policy and relations. I realized nuclear fusion stopped getting funded at some point and I was like, “Well, you have to have both skillsets.”
Sarah Nooravi: Anyways, I ended up finishing up my master’s.
Kirill Eremenko: Wait, just hold on. Sorry, so you’re not going to tell us that hilarious story about how you chose mechanical engineering? We’re not letting you off the hook here.
Sarah Nooravi: I mean because look, when you get into, when you realize you want to study engineering, that’s part of the battle. Like, “Okay, now I know I want to study engineering.” Then you realize, so I went to UCLA campus and I was like, “I want to study engineering!” They were like, “That’s cool, what engineering?” I was like, “What do you mean? How many engineerings are there?” They’re like, “Well … ”
Kirill Eremenko: I know, right? I didn’t know there’s like civil, mechanical.
Sarah Nooravi: Electrical, mechanical.
Kirill Eremenko: Chemical, any kind of thing engineering. Just like put a noun and then engineering after it, it exists. It’s crazy.
Sarah Nooravi: Yes, exactly. I was like, “Oh, okay, yeah, let me go back and think about it.” Then I was like, “Okay, maybe it was civil.” I didn’t know how I decided on civil, I was like, “Civil.” Then I went to the civil department and I was like, “Oh, so I want to apply to the master’s program here.” They were like, “Oh that’s great, what specialization?” I was like, “Excuse me? What are you talking about?”
Kirill Eremenko: It keeps going.
Sarah Nooravi: At that point I realized I don’t need to go the top down approach, I need to go the bottom up approach. I need to figure out what exactly am I trying to specialize in, whose research am I excited about and then I can decide, I can back out. “Okay, well oh, that was mechanical the whole time,” you know? Because I found a professor that, I loved her research. It was on solar powered power plants and renewable energy storage and I was like, “Okay, this is exciting, I want to do this.” I met her in a parking lot. I talked to her when she had a flat tire. I was like [crosstalk 00:22:14] annoying.
Kirill Eremenko: Tell us, Sarah, how did she get a flat tire? Did you happen to do anything, have to do anything with that flat tire?
Sarah Nooravi: No, right? At just the right moment. [inaudible 00:22:26], no, but that’s the thing, right? If you’re excited and passionate about something, and think about me. I never had any experience in engineering at all. Just from my story you can tell how junior I was. You see this type of, the same thing going on with people trying to get into Data Science. It’s this desire of like, “Oh my God, I see what I want and then how do I get there?” You have to be kind of scrappy. Like who are the right people that you need to connect with and talk to and show them that you’re passionate and meet them in a parking lot when they have a flat tire and just go out of your way to make things happen for yourself. You have to really be ready to put in that type of effort and be gritty to go after it.
Kirill Eremenko: Gotcha, yeah. Totally agree, totally agree. The best part is what I love about the way the world works is when you really like that and you really, truly want something, things will happen to align in your favor. Flat tires will happen just at the right time when you’re walking past the car park. Things like that.
Sarah Nooravi: Yeah, yeah.
Kirill Eremenko: Okay, cool. You picked a professor whose research you liked. I just didn’t realize that solar was part of mechanical engineering.
Sarah Nooravi: It is, yeah.
Kirill Eremenko: Interesting.
Sarah Nooravi: All of the renewable energy type projects fell under mechanical and so I specialized in heat and mass transfer which is essentially what all the thermodynamics is doing.
Kirill Eremenko: Okay, gotcha. You became a researcher at UCLA in the mechanical engineering space. Is that correct?
Sarah Nooravi: After that, I mean so then here we get to the point of so many people, of the job market. I now am graduating and I’m approaching the job market and I’m like, “Okay, so I have an undergrad that’s focused in economics and math and then I have a graduate degree in mechanical engineering.” I was like, “You know what I’m going to do, I’m just going to create a resume for both and the job that I get first will be the direction I end up going.”
Kirill Eremenko: Interesting.
Sarah Nooravi: I left it up to chance and the job market to dictate where I ended up and fell in love with the culture at a startup in Hollywood. I just loved the culture, I saw myself fitting in there, I liked my manager, I liked the projects that they were working on, the direction the company was moving. It was very inviting to someone like me. I don’t want to say that it happened by accident but I didn’t go out searching for it. It just kind of was like leveraging whatever skillsets I had and then from there, yeah, I don’t know.
Kirill Eremenko: Wow, you did such a good job at keeping it, not telling us. I’m sitting here dying to know which one was it, was it the mechanical engineer or the mathematics? Which one did chance pick for you? What was the startup involved in?
Sarah Nooravi: Yeah, yeah, oh sorry. No, I ended up doing analytics.
Kirill Eremenko: Okay.
Sarah Nooravi: Yeah. I actually never worked a day in my life as a mechanical engineer. I studied it and I thought, “You know, maybe eventually it will come in handy.” I know there’s a lot of companies right now in energy that are going towards IOT, all the smart grid and stuff like that. I think that’s what I would have liked to do but I think emphasizing more on the data side right now could actually be leveraged eventually into that industry anyway.
Sarah Nooravi: Yeah, I started working as an analyst, data scientist, picked up and filled the gaps of all of my knowledge with the Machine Learning stuff and the Data Visualization and et cetera, et cetera, and then we get to where we are now.
Kirill Eremenko: Was it hard to pick up all that knowledge, the Data Visualization and Machine Learning? How long did that take you and was it a chore or was it more of an exciting path?
Sarah Nooravi: That’s an interesting question. For me, graduating with a minor in stats and then I studied economics as well, they went over a lot of the fundamentals. Your linear regression, your logistic regression, dealing with literally every, I took at least two years of really understanding that stuff very well. Then you go into a company now that’s focused on predictions and predictive analytics and you realize, “Oh wow, I just was not prepped for this at all.” We didn’t learn Machine Learning as a part of our curriculum.
Sarah Nooravi: Seeing that the company was gearing itself in that direction, I was like, “Wow, I really have a lot to learn.” My way of learning, and people who know me or interact with me locally, they know that the way to learn, at least for me, is to teach. I started teaching myself all, filling the gaps of all the things I needed to know and then hosting monthly Machine Learning meetups in LA where I would just basically talk about what I learned that month or like a project that I was working on that maybe people would be interested in hearing about. I just took that on myself to make basically, and we go back to Rico, oh Rico. You will forever be remembered for commit, fail, improve.
Sarah Nooravi: I just by committing myself every month to a meetup that I had to get in front of people and talk to them about something Machine Learning related was my way of just holding myself accountable for learning and then also integrating my learnings and my conversations into the projects that I was doing. It definitely wasn’t overnight and I’m still learning, so what’s great about being in this space as well is that you’ll never learn everything.
Kirill Eremenko: Yeah.
Sarah Nooravi: You can get proficient, you can be very good at several things and then know of many things but you’ll never know everything.
Kirill Eremenko: Yeah. I agree. The way that, I love that approach. As you said, Rico with his reckless commitment, that is so cool. To learn something, you commit to hosting or explaining it at a meetup and that forces you to learn. Tell us, did that go well every time or were there times when you found that the challenge was too complex and you just couldn’t possibly learn it on time?
Sarah Nooravi: That’s a good question. For me, I always went with topics that were aligned with projects I was working on. In the months that I knew I couldn’t pull it off, I delegated. I chose a victim from my company to do the meetups.
Kirill Eremenko: Nice.
Sarah Nooravi: Then it works out, right? Because I think that consistency, because what I was trying to do was at the same time as learn myself was develop a community of people who could rely on each other and feel like that they were being supported. I feel like there was such a demand for it in LA every time that I held a meetup. I mean the first one I did was in a coffee shop. It was at Coffee Bean, there was like maybe 30 people that showed up and I didn’t know anything at that time. I was like, “Oh, let’s just pull up,” Sklearn has their nice diagram of all the different models. I was like, “All right guys, let’s just pull up a model and then go learn it for 30 minutes and then come back and explain it to everyone.” That was my first meetup.
Kirill Eremenko: Wow. Wow, that’s crazy. 30 people sitting there staring and you’re like, “Okay.” That’s so interesting, wow. You do this through Meetup.com?
Sarah Nooravi: Yes I do.
Kirill Eremenko: Okay, wow, and you still do it to this day?
Sarah Nooravi: Til this day, yeah. I moved out of LA, and to answer your question, yeah, I did feel like sometimes it was a struggle but I think having that commitment, like, “Oh, 100 or maybe 100 people, 60 to 100 people are relying on me to follow through on this [inaudible 00:31:39]. I’d better have something good for them.”
Kirill Eremenko: Yeah, yeah.
Sarah Nooravi: Once I moved out of LA and I moved to Irvine, MobilityWare has been very, very gracious with allowing me to kind of keep that going and providing us with pizza and a space and just all the accommodations. It’s been very, I’ve been very fortunate with not having to worry about a venue in order to host these and keep them consistent and build a good community. Next year, I think I might have help from one of our fellow Data Science influencers to help me keep the LA chapter and the Irvine chapter open and expand the meetup and keep it going.
Kirill Eremenko: Nice, nice. Who’s that? Who’s the influencer if you don’t mind disclosing?
Sarah Nooravi: It’s Randy.
Kirill Eremenko: Randy. Randy’s going to, oh, that’s awesome. Randy’s great, that’s so cool. People love him. That’s awesome. That’s so exciting for you. MobilityWare is the company where you work currently, is that correct?
Sarah Nooravi: Yeah.
Kirill Eremenko: Awesome. Tell us a bit about, what do you do there? What’s your role? Because there’s so many different ways companies use Data Science these days.
Sarah Nooravi: Right now I’m the sole dedicated marketing analyst. I do everything for our marketing team. My job is a little bit, it’s more than a full time job I would say because I work across all of our games and we have three different suites of games. Our card suite, our casino suite, social casino, and then we have puzzle. Each of them, it’s very interesting, they’re all in different stages of their life cycles so some of them are just starting out and we’re trying to prove whether or not we need to continue to sustain them and do UA for them or they’re pretty much stable. Like our solitaire game, it’s been basically there for a very long time and so the marketing strategies around our different games are very different.
Kirill Eremenko: Sorry, I missed that. These are games for mobile phones, right?
Sarah Nooravi: Yeah.
Kirill Eremenko: Okay, gotcha.
Sarah Nooravi: My job entails really surfacing data to our marketing team because before I was here, I think that was a part of the struggle. I touch every database that we own and consolidate and really surface that to our teams so that they can make better decisions, but then aside from that I’m building out a lot of tools for them. Like competitive benchmarking tools, creative optimization tools, different campaign optimization tools which will all be initiatives that I’ll be running next year but I’m also working on now. Sometimes I get pulled into things to do with product, so understanding user behavior, developing user segmentation models. I kind of get to touch everything which is nice about my role.
Kirill Eremenko: That’s cool. You mentioned you’re the sole marketing analyst in the company.
Sarah Nooravi: [crosstalk 00:34:51].
Kirill Eremenko: Yeah, you were mentioning that as well before the podcast, that you’re thinking of expanding the team as well. Tell us a bit about that. Like when, because I also went through a similar situation where I, after Deloitte I joined a company and I was the only data scientist for a while. I’d be interested to hear your experience. At what point do you realize that this or the company realizes that this is beneficial, that there is value in having a data scientist on board, let’s start growing the team? What are your thoughts on that?
Sarah Nooravi: I really think that it depends on the company and who’s at the top and whether or not they see, the reason why I say that is because I’m thinking about two different scenarios. In one scenario, you build out tools and you basically prove your value through those tools. It’s like, “I can show you that revenue is going up because of these models that I’m building and the campaigns that we’re doing, the experiments that we’re doing and through basically having data scientists analyzing the data, building models, et cetera.” If it’s very clear, like, “Oh, I can see the revenue lift as a result of having a data scientist on board,” then you don’t really have much argument there.
Sarah Nooravi: In other companies, maybe it’s a little bit harder to justify when there’s no real, you can’t point to revenue and say, “Hey, our revenue’s increasing because I exist.” From there maybe it’s a little bit of a harder discussion to have but whether you can prove that through automation or optimizing what domain experts are doing and helping them do their jobs better, that’s a way to do it. I think on my end, I’ll speak to my current job, the tools that I’ve build out for marketing have just been amazing. Their words.
Kirill Eremenko: That’s awesome.
Sarah Nooravi: They’ve really appreciated having someone dedicated to their needs and especially since we have a lot of budget allocated towards UA and marketing in general, the initiatives that I want to run next year are just too much for me to handle alone. I’ve kind of pushed for maybe having someone on my team or having a few people on my team that we can all work towards driving better decision making on that side.
Kirill Eremenko: Okay, okay, gotcha. Interesting. It’s ultimately up to the data scientist to show the business value, to make the case, to make it a no brainer decision for the business to go ahead, right?
Sarah Nooravi: Yep, yep.
Kirill Eremenko: Okay, okay, makes sense. Cool. Can you give us an example? It’s a very interesting industry, I don’t think anyone on the podcast has been before talking about games and mobile phones and it’s a massive, massive industry. There’s a lot of games popping up all the time for mobile phones. What is like a recent project that you are proud of and that you’re able to share with us some details, maybe some tool that you used or some approach or some, kind of like more industry specific use case of Data Science that you can tell us about? Is there anything that comes to mind?
Sarah Nooravi: Sure. I have two in mind but maybe I could speak to the one that just got productionalized recently, but it doesn’t deal with marketing. It’s more on the product side.
Kirill Eremenko: Sounds good. Give us a little insight into this world. That would be very cool.
Sarah Nooravi: On the product side we have a lot of users who come into our game and have some certain user behaviors. For us, what we can do or what we’re aiming towards is as much personalization within the game as possible because on our side we want to create a good user experience and eventually some sort of purchase or some sort of engagement so it’s a win win. For us, I think one thing that we were hoping to do was really understand our users in terms of different segments. I mean most of our users or listeners might know of K-means, so doing a clustering model on our user segments. Even though K-means isn’t hard, really understanding what, so the upfront on this is really understanding what features really needed to play into differentiating these different segments in order to create really good, well defined segments that we can now create campaigns off of and develop these personalized store configurations or messaging in order to create a better user experience.
Sarah Nooravi: The reason why I’m proud of that is because we just closed the loop in our data pipeline, so not that this model doesn’t just exist on its own. This is something that I talked about in a recent article that I wrote, which is that most businesses are suffering from the cold start of AI. They don’t have that closed loop of, whether it’s the data infrastructure or whether it’s taking the model output and actually using it. What I’m excited about is the productionalization of my model which is now taking the output and it’s pushing to a live environment where we can actually build these campaigns and do something with the model output rather than it sitting in a PowerPoint or sitting on Jupyter Notebooks or in a Python script somewhere.
Kirill Eremenko: That’s really cool. You’re right. Productization of Data Science outputs, it’s a whole new world. We often think, “Okay, I’ve got the insights, I’ve done the modeling, I’ve got the insights, here’s the presentation, done.” No, that needs to go to the IT department or whoever else and that needs to be implemented, like it might be actually you might have to reprogram it in a different language. You might have to create some sort of protocols for it to talk to the existing servers and infrastructure and it has to somehow be integrated. It has to have its own window during the night when it will be running. How often does it have to refresh? How do you maintain it? Who looks at the results? How do they get integrated in the company? That’s like a whole new project on its own.
Kirill Eremenko: Tell us a bit about, first of all, congratulations. That’s a massive win, but it would be really cool to hear like were you involved in, how did you hand over this part from, like you created this K-means cluster algorithm which I think actually a pretty cool approach to creating a better user experience. Let’s cluster our users and find out what kind of groups do we have, but then how were you involved and what is the process like of taking what you create and handing it over to the people responsible for productization of your development?
Sarah Nooravi: Yeah. I had to work really closely with our engineering team who were specifically building out this process for us to essentially schedule the output. My script runs every day and I had to work with them to figure out, okay, so they came up with a wrapper that will essentially take the output that I, the script that I’m running and it’ll wrap it within the activation process that they have. Then it’ll push to a live environment and so I had to work with them to understand a lot of what GitFlow is.
Sarah Nooravi: I know how to use Git and I know very basics of committing but GitFlow is a whole new world of taking you through different environments. From dev to test to stage to prod and really having them walk me through that and working really closely with them so that, when you’re pushing to a live environment you don’t want to break anything. You want to make sure that you’re testing every step of the way and you’re doing QA on your output every step of the way. Learning that process, working really closely with the engineers to help document that process so that we can, the next time that we want to work on a project like this or a productionalized output, that it’s stable and that it’s easy to follow.
Sarah Nooravi: What else? I think mostly the hardest part was really getting, because I was one of the first people to help productionalize output through that process, so it’s really like understanding how it is so that I can eventually teach our team. Then working, so that was the engineering side of just getting it in team with stakeholders. The person who, the PM, the product manager for that particular game, I had to work with them on developing, “Okay, well what’s the attribute going to be? How do you want it labeled? When you call it in your live environment,” all the nitty gritty of what they need from their side.
Kirill Eremenko: Yeah, wow. That sounds like an involved job. How long did the project take you and how long did the productization take you?
Sarah Nooravi: The project took me, I would say from scoping out the project to actually implementing and being done with the model, maybe a couple of weeks. Then the productionalization of it, also probably a couple of weeks. In total probably around a month or so.
Kirill Eremenko: Interesting. The productization takes as much as the project itself.
Sarah Nooravi: I think it was a little bit slower only because I was learning the GitFlow process and really on the, it was very engineering heavy. Being the first one doing it, obviously there’s hurdles. I think second, third time around, that process will be much quicker.
Kirill Eremenko: If you don’t mind sharing, why did you pick K-means clustering out of all the available algorithms?
Sarah Nooravi: That’s also a good question. I was familiar with it and it’s simple, there’s really nothing too complicated about it. I think because I was familiar with it and because I needed to have such quick turnaround it was a project that I, it didn’t necessarily, it wasn’t my highest priority but it was a priority. I was like, “Okay, can I get good results using K-means?” When I saw that it was performing pretty well and I was getting results that seemed reasonable and that I could put into effect pretty quickly, I was like, “We’re just going to run with it.”
Kirill Eremenko: Gotcha. That’s the way to go sometimes, right?
Sarah Nooravi: Yeah.
Kirill Eremenko: It’s fast, you get results, the 80/20 rule. Why would you spend, you already spent in total like a month on this project with the productization, why would you spend six months on it if you can already implement something and get the results? That’s very cool, that’s very cool.
Kirill Eremenko: Okay, well thank you very much. That’s a very interesting case study. I’m sure a lot of our listeners got a great insight into this world and especially this whole productization approach. I want to switch gears a little bit and talk about something that I think we’re both passionate about and that is diversity. When you were at DataScienceGO, you spoke on the panel of Women in Data Science and we had an interesting chat about in general how to enable, empower more women to get into this space just before the podcast. I would love for you to share your thoughts on this with our listeners if you don’t mind.
Sarah Nooravi: On the importance of diversity?
Kirill Eremenko: Yes, please.
Sarah Nooravi: Sure.
Kirill Eremenko: Importance of diversity and what can we do as a community of data scientists to help anybody regardless of their agenda or ethnicity, background, to be able to get into this space and really benefit not just like an individual company but the community in general and bring those new ideas, fresh perspectives, insights into this community that we’re building of data scientists.
Sarah Nooravi: Sure. In terms of the importance, I think every company that wants to maximize the production that it’s making within its business and get the best ideas to come out and the best solutions to the problems that they’re trying to solve would think about diversity as one of the key factors that it would need to try and incorporate. This has been proven time and time again, where diverse teams will outperform non-diverse teams on different approaches and solutions to problems. Especially when we’re dealing with like challenging and complicated problems that involve the entire world at this point, we’re trying to make solutions that affect everyone, having a team that looks representative of who they expect their consumers to be would be important.
Sarah Nooravi: The one example that sticks in my mind and it’s like forever since I heard it was a woman who worked at Google X when they were testing out their Google Glasses. She was like, “You know, I’m on the panel to essentially test out the product and then come back a week later with feedback.” She was like, “Yeah, so I took the Glasses, I wore them for a week and then I came back to talk to the team about my feedback,” and her feedback was essentially like, “When I take the Glasses off, it sticks to my hair, so hair pulls out when I take it off.” The guys in the room were like, “Well, why don’t you put your hair up?”
Sarah Nooravi: It’s like, wait, hold on, do we really think that that’s the solution?
Kirill Eremenko: That’s so funny, that’s so funny.
Sarah Nooravi: When you think about putting together a team and really creating products and solutions for the masses, you have to have a team of people and be open minded and hear that feedback but be actually willing to do something about it.
Sarah Nooravi: The importance part, I don’t know that I need to argue too much. I know that we can all agree that diversity is important, right?
Kirill Eremenko: Yeah.
Sarah Nooravi: Why it’s challenging is that it’s not very common yet. Whether it’s the minorities or the women or even minorities of educational background. Maybe someone who studied humanities who wants to get into Data Science or someone just doing something totally random that you wouldn’t expect who has an analytical focus and they want to get into it and this imposter syndrome that we talk about. I think everyone can share in this idea of that we’re trying to figure out where exactly we fit in but by embracing our differences and by being okay with, “Hey, you have a different perspective than I do and that’s okay.”
Sarah Nooravi: The reason why diversity helps is because when you think about it, when you get into a room and you see everyone that looks like you, you don’t think that you need to press your point too much. You assume that everyone thinks like you, but when you enter a room and people don’t look like you and you’re like, “Wait,” or you know that they come from different backgrounds, you’re like, “Okay, I need to convince people of my point.” That’s why the diverse teams work, is because everyone’s now talking about and actually expressing their perspectives and now a discussion gets made about it and then you arrive at the best solution.
Sarah Nooravi: Within our community, I think what we can do is understand that that’s our goal. Our goal is always the same, right? We’re always aiming towards the same goal, that we want to achieve the best product, we want to build an inclusive community and a lot of that is just embracing someone else’s differences. Being like, I know this is going to be hard. Diversity doesn’t just happen overnight and it doesn’t take no effort. It’s hard to accept someone who looks like you thinks differently from you, et cetera, et cetera.
Sarah Nooravi: Things that I’ve done that I think that other people can do is just start these conversations. Tell someone that they did well. Reassurance like, “Hey, you did really well on that, that was really great.” Make people feel validated in what they’re doing, like as managers or as colleagues or as friends. “Hey, you did really well on that, that was really impressive.” Positive affirmations could go a long way. Mentorship can go a long way. Standing up for someone when you feel like they have no voice. Like sometimes depending on who’s speaking in a room, you may listen to them differently and so giving someone who otherwise maybe wouldn’t step up and defend themselves, “Hey so and so, you had a very good point about X, Y, Z, do you want to talk about it?” Like helping support each other within teams and within the community could go a long ways.
Sarah Nooravi: I did want to mention, something that I’m doing is I started a mentorship program called GLAD. It stands for something hilarious. My creative team, some guy on my creative team came up with the name. It stands for Glamorous Ladies and Data.
Kirill Eremenko: That’s nice. That’s really smart.
Sarah Nooravi: You know, it’s funny but I don’t want it to be geared just towards women, right?
Kirill Eremenko: Yeah.
Sarah Nooravi: I want it to be a very inclusive environment where we can just work together and develop a supportive community. It’s essentially just that. It’s bringing people together and just building up confidence and reassuring people and helping them through a lot of, sorry, I’m going to go on a tangent but a lot of Data Science is, it’s the technical side but it’s also a very emotional journey.
Kirill Eremenko: Yeah, yeah, yeah, totally.
Sarah Nooravi: When you realize you’re helping someone through their journey into Data Science, it’s not just, “Hey, let me help you find the best models or let me help you with resources.” It’s also a lot of reassurance. It’s that emotional side of, “Oh, the imposter syndrome, do I feel like I belong here? Maybe I’m not qualified, maybe this isn’t what I should be doing?” It’s really helping people build that confidence regardless of who they are, right?
Kirill Eremenko: Yeah, yeah, totally. Totally agree with you. Thank you very much for that very inspiring talk and also good suggestions. Positive affirmations, mentorships, standing up for someone, even just saying, “Hey, what do you think about this?” I want to add to that what we talked about just before the podcast were, role models. Role models are super important and the whole, it’s really hard for somebody from like, for instance, a woman to get into Data Science when they don’t see that many data scientists.
Sarah Nooravi: Yeah.
Kirill Eremenko: When they see, women see that there’s only 10% on average of data scientists that are female, then that’s what you will get in terms of people entering the field. We want to improve that and so the best way to do this is to show that there are actually lots of successful women who are enjoying being in the field of Data Science. That doesn’t mean you have to be like the best data scientist in the world. All you have to do is just show up. Go to a meetup and be present and that will show people or show women who want to get into Data Science, will that you are there, you’re a successful woman in Data Science. Or like maybe invite somebody to a talk or try to present at a talk. Things like that. Just more publicity in that space for women will attract other women into this space. That’s kind of my thoughts on how we can help in the sense of role models.
Sarah Nooravi: 100% agree, yeah.
Kirill Eremenko: Awesome. Okay, well Sarah, thank you so much. We’ve come to our time limit on this show. Thank you so much for coming and sharing all these insights, totally loved the chat. It was lots of fun exploring your background. Before I let you go, what would you say are the best places for our listeners to get in touch with you, contact you and follow you and your interesting career, see what you get into in the years to come?
Sarah Nooravi: I think LinkedIn has definitely been that one platform where we’re developing all of our network in Data Science. Linkedin is probably the best way. I’m not really active on Twitter yet.
Kirill Eremenko: Gotcha.
Sarah Nooravi: [inaudible 00:59:54] when I do.
Kirill Eremenko: Gotcha. Awesome, awesome. Yeah, you guys, so listeners on the podcast, Sarah has 23,000 followers so make sure to join all the people benefiting from the things you’re sharing. By the way Sarah, I had a look at your recent article, Creativity in Data Science, very, very interesting. I also like the talk by Sir Ken Robinson on TED and I like how you incorporated his ideas into Data Science and the whole notion about creativity. I highly recommend for others to check that out as well.
Sarah Nooravi: Thank you.
Kirill Eremenko: Okay, and I have one final question for you. What is a book that you can recommend to our listeners to help empower their careers?
Sarah Nooravi: Okay, so I am reading this book currently. It’s called, which you’ve probably read it, How to Win Friends and Influence People by Dale Carnegie. You’ve read it, right?
Kirill Eremenko: Yeah, amazing book.
Sarah Nooravi: I think that after reading that book, and I just put up a post not that long ago, maybe last week, talking about the importance of human relations. Especially me who’s thinking about the future and where things are headed, I think understanding how to deal with people and especially when you get up in the ranks as a manager or a director and you’re dealing more with the human side and less on the very technical nitty gritty side and especially with things getting automated the way they are, I think that people should be looking towards improving their communication skills and looking towards how do they improve the relations with people and their human skills.
Sarah Nooravi: That’s one that I would say, if you ask me next week maybe I’ll have a different book but I think if you’re thinking holistically about the Data Science space and of different skills that you should have and you think about a well rounded data scientist, I think this is definitely a book that everyone should read.
Kirill Eremenko: Fantastic, and that book can help you not just in Data Science but in all other aspects of life as well.
Sarah Nooravi: Exactly.
Kirill Eremenko: Awesome. Thank you so much, Sarah. Amazing having you on the show today and really appreciate you coming on and sharing all those wonderful insights.
Sarah Nooravi: Thanks, I had such a great time.
Kirill Eremenko: There you have it. That was Sarah Nooravi. I hope you enjoyed this episode as much as I did and my personal favorite part was how open Sarah was, how positive this episode turned out and how many laughs we had. That was very exciting, very fun, and you can tell right away that most likely Sarah is extremely successful in presentation skills and communication. No wonder Sarah recommended the book How to Win Friends and Influence People by Dale Carnegie. I think we all as humans can pick up some interesting tips and ideas from that.
Kirill Eremenko: As always, you can find all of the show notes for this episode at www.SuperDataScience.com/227. That’s SuperDataScience.com/227. There you’ll find all of the materials that we’ve mentioned on the show, including the URL to Sarah’s LinkedIn. Make sure to connect, make sure to follow Sarah and get all these interesting updates and insights that she’ll be sharing in the near future. Make sure to forward this episode to somebody you care about and somebody you want to inspire. On that note, I look forward to seeing you here next time. Until then, happy analyzing.
Show All

Share on

Related Podcasts