Why you should not be a data scientist

hey guys how's it going i've been a data scientist in tech for a little over a year now and while that's not a very long time i have had the amazing opportunity working with very very skilled data scientists people who are far more senior than me i've had the chance to shadow and to learn from them and also be mentored by them and recently i've also started mentoring more junior data scientists than myself as well this has really got me thinking about the role of a data scientist i believe in a growth mindset which means that i think your talents and your skills are not static if you want to become a data scientist you can work hard and you can work towards that goal and become one however with that being said i've really been reflecting and thinking about what is the role of a data scientist and also kind of wondering if people actually want to become a data scientist like do you want to become a data scientist because you really want to become a data scientist or do you want to become a data scientist because the pay is good um you get to work in tech it was named the sexiest job in like two years ago or something like that so in this video i want to talk about why you should not become a data scientist number one is if you're not motivated to learn a hybrid blend of computer science statistics and business data science is a very interdisciplinary field you have the computer science part you have the statistics part and you have the business part at the minimum so i see a lot of people who start off in computer science or business or statistics and maybe you have a combination of those two but it's very rare for someone just to you know come out of school and just have all three immediately which means that you're gonna have to take what you have in that background that you have and probably expand to the other two for me specifically in terms of traditional training i come from a computer science background although i have a little bit of statistic knowledge because i did a science pharmacology undergraduate but my major training it was in computer science which meant that i picked up on a lot of statistics and especially on the business side a lot during my role people have asked me i come from like x background can i be a data scientist and the answer is always yes you can be a data scientist however i did notice that if you come from a more technical background say from computer science or from statistics and math it's generally easier for you to pick up the business part of it the the non-technical portions of it of course i have also seen people who come from a business background and they're amazing data scientists right now some of my co-workers are but it did usually take longer for them and it was more effort for them to go and then learn the computer science um to learn the coding components as well as math components if you're someone who's very interested in one of these fields and just not really interested in picking up the other two for example then data science may not be the best career for you truth be told learning all of these other interdisciplinary fields isn't super easy you know it's not something that you just spend a weekend on it's actually a very gradual continuous progress and you need to be motivated to do so and also disciplined to do so and filling in the gaps and it often entails asking really dumb questions sometimes for me because business was the part that i had the least training in going to meetings that were business related and asking like really dumb questions like what does this acronym mean for example it also means admitting your mistakes you know when you write code that produces a result that is incorrect then owning up to that mistake saying i'm very sorry and then fixing it um and then next time making sure you that you don't do it again and finally also just picking up these skill sets constantly that your mindset isn't like okay i'm only gonna learn like these new things today and tomorrow and then spend the rest of the week not learning anything it's it's always like a gradual process of picking up more and more bits from here and there learning from other people learning from your mentors both formally and informally if any of these things sound kind of not up your alley then becoming a data scientist may not be what you actually want second reason for why you should not become a data scientist is if you don't feel comfortable being scrappy so what do i mean by scrappy it means not being super focused and wanting to produce an amazing product it means doing whatever it takes using whatever tools that you need in order to accomplish the goal here's a scenario that has come up for me many times someone comes up to me and goes like tina we need to do this super duper important thing and we need it to be done in three days i'm just like there's no way i can do this thing in three days if i do it you know like properly write all like the correct code and you know just get everything like perfectly tested making sure that everything is functional there's just no way right so what i have to do is i have to prioritize i'm like okay i could do this thing in three days what's the best way for me to do this so that i can cover most of the things that this person needs me to do um how can i be as accurate as i possibly can you know not not accidentally make any mistakes that could cost the company a lot of money and that usually means being really scrappy and putting together a bunch of different combinations of tools it could be like some sql code that's often packaged around like some python code also just doing some math calculations by hand sometimes just inputting that in and n is just scraping together all of these things and producing an mvp or a minimal viable product for whatever it is that needs to be done it's imprecise it's not perfect the code sometimes hurts my soul but it gets the job done and i understand that as a data scientist that is oftentimes what i'm going to be doing because data scientists generally have such a diverse skill set in business statistics and computer science it also means that i often take on the role of roles that are not data science for example if there's a bunch of data that's there that's not clean and we need to write pipelines about it and i become a data engineer sometimes you need a ui component where you need a front end then you kind of become a mini software engineer sometimes we need to present a strategy to leadership to convince them that this is the right approach then you kind of become a hybrid business person product manager i've personally done all of these things in addition to what my core role is as a data scientist if you're someone that prefers to focus on a single project and doing things in a very specific way then data science may not be the best job for you number three is if you don't like learning new things constantly remember what i was saying earlier how you have to learn computer science you gotta learn business and you learn statistics and math maybe you're like all right i'm gonna go learn all these things and then i'm gonna become a data scientist and everything will be wonderful right wrong wrong wrong data science in itself is such a new field and it's progressing so quickly and just morphing in front of our eyes two years ago what data science were doing is really different from what data scientists are doing today and data scientists of tomorrow it's going to be really different from what we're doing today without even exaggerating it feels like every week there's a new tool or a new way of doing things that's coming out and as a data scientist it's your job to solve problems efficiently and effectively and what that means is that you need to take that initiative to keep up skilling yourself learning these new tools that are coming out first of all actually figuring out if these tools are worth learning about because there's way too many things to learn about if you try to learn everything so figuring out what's actually important and then going and actually up skilling yourself to become better and just keeping up with that industry as a whole to survive you really need to be constantly learning and from my own observations of much better and more senior data scientists than myself i notice that they really do two things the first one is that they're always taking on projects that are outside their comfort zone and it's like that accountability because it forces you to go and learn new tools in order to expand your scope and you become better and the second part is that they consciously put aside time to learn new technologies even if it doesn't have to do with a current project that they're working on because they just want to expand their toolkit at some point they may be able to use this new skill set to solve a problem better or faster or solve a problem that they previously were not able to do so just that constant learning that's going to be going on throughout your career is not something that you're interested in doing and it doesn't really resonate with you then you probably shouldn't become a data scientist fourth reason why you shouldn't become a data scientist is if you don't embrace the scientific method theta scientist notice that second word scientists it's actually very very crucial the word scientist is there and that's because data scientists are first and foremost scientists and a scientist is someone that follows the scientific method and what that entails is first doing some research to figure out what your problem what your question is and then forming a hypothesis and then the next step is taking your hypothesis and conducting some experiments we're creating some models we're in some way trying to answer your hypothesis to see if it's correct or not and then from these results you would then draw some conclusions or you know decide that you need to work more on it and then ultimately present those findings which hopefully will be able to influence strategy or improve the company in some fashion and in some way in a nutshell hypothesis-driven decision-making data scientists are the equivalent of harbingers or truth i like to call that publisher is like a really weird word but it's like the you are the person that kind of has to hold people accountable to the truth the truth of what the current state of affairs actually are when you make decisions where when you make recommendations it has to be grounded in the truth you're using the tools of computer science business and statistics in order to inform your decision in order to build something like forecast more models so you can't just be like uh yolo i think we should do this because my spidey senses are tingling where like i feel it in my soul again your decisions and your recommendations and models you build have to be grounded in data they have to be grounded in truth with that being said though i'm not saying that data science isn't a creative field because it definitely is you know science is always going to be half our half science but you do need to be someone that really respects the truth you really are building upon the truth and you embrace that because if you don't keep everyone honest nobody will alright fifth reason for why you should not become a data scientist is if you don't like marketing your own work i know i know the word marketing it sounds it sounds terrible right like ideally and what i thought it would be like would be like i would do this really cool analysis and then it's gonna be so awesome and i'm gonna show it to people and they're gonna automatically think it's amazing and they're going to go and implement it yeah no sadly that doesn't happen even if you have an amazing analysis you have an amazing thing that you discover or amazing model that you built you actually have to self-advocate and do some marketing in order to get people to use these insights that you found out to actually use the model that you built so what you need to do is of course do all the amazing data science stuff go build your models to go do your forecast go and figure out your strategic insights but then after you do that you also have to work on presentation you have to work on understanding what the business needs how to go to someone and be like this thing that i made is going to make this company better this thing that i made is going to help you achieve your goals because of xyz reasons where else if they don't understand why what you did is valuable they're not going to use it so whatever it is that you did is never actually going to see the day of life this is this is what i found the really strong data scientists that i've seen they not only do amazing data science but they're also it's really good speakers really good communicators through a lot of different forms through communication verbal communication through written communication um just through like being able to network with people understand what their issues are and just having a really deep understanding of the product and the space that they're working in to make sure that the work that they're doing is extremely valuable to that space so if you're someone that prefers to just really focus on the technical components and just let the work speak for itself data science may not be the best field for you all right we've come to the end of this video i just want to make a final disclaimer by saying that this of course is just my experiences i am 100 sure that if you work as a data scientist in different companies it's going to be very very different especially because data science is such a new field you're going to have different varieties or flavors of data science and a lot of different industries and in some industries it's going to be vastly different from what i've described all that i can say is that this has been my experience as a data scientist specifically working in a big tech company so do take what i said with a grain of salt but i do think these different reasons for why you you may not want to be a data scientist are non-obvious things that people don't really think about because data science is such a hyped up field that has good pay you know that has a lot of media coverage i do think a lot of people come into this field expecting something and not actually realizing what data science is which makes them not very happy as a data scientist and they ultimately may have you know done data science for a couple years and transitioned to another role so in order to save you that if you are considering being a data scientist do think about that are these things that resonate with you do any of these stand out as something that you just really are not interested in and if so you know really deeply consider if you want to be a data scientist if you actually want to do this or not alright thank you all so much for watching i hope this has been a helpful video for you guys and i'll see you guys in the next video or live stream

Channel: Tina Huang

Views: 718,678

Keywords: data scientist, data science, don't be a data scientist, you won't be a data scientist, how to be a data scientist, machine learning, why you should not be a data scientist, tech data science, facebook data scientist, amazon data scientist, apple data scientist, netflix data scientist, google data scientist, tina huang

Id: sOZ8MxFw8TQ

Length: 12min 32sec (752 seconds)

Published: Sun Sep 19 2021