Five hard truths about building a career in Data Science

Key Takeaways

[Music] hello so a simple google search on how to build a career in data science will throw a long and exhaustive list of skills in your face you know from learning how to program in python then learning statistics mathematics and you know finding out how to make reports dashboards using tableau or excel or any other tool then being proficient in communications and yeah there are n number of lists out there that you need to master in order to become a good data scientist now to add to this chaos you have a very large spectrum of job profiles for example uh you know a fintech company might require a phd in data science or phd in finance or mathematics in order to build their algorithmic trading strategies whereas a product-based company might just need a student statistician who can run and design their a b testing experiments to check which features are working which features are liked by the users which features would work well so as a result of this whole list and you know this large landscape of data science and the skills that are required to master as parents who are still trying to figure out whether they want to build a career in data science it becomes very overwhelming for those now in this video i am going to break down these hard truths that you should be aware of before you get down to building a career out of data science so five hard truths that you should be aware of before you start making a career in data science so truth number one nobody can do everything you should look to become data literate and then narrow down there's not even a single data scientist or data analyst who knows it all who possesses all the skills that we have just mentioned so what you should do is you should look to become data literate which is basically you should acquire enough mathematical and statistical knowledge so that you can understand your data understand the problem understand the whole frame of your reference your context and everything and then you should look to specialize in two to three skills that's it data scientists who do good they specialize in two to three domains or and two to three skills on top of their basic data literacy which make them irreplaceable and what you should do in order to become data literate is basically expand your knowledge of available methods and techniques so if you do not know something if you do not know a technique how would you apply it first of all so that is expanding your knowledge second choose the right method so you should be able to assess which method or which technique you should apply and you should be able to assess that for your particular scenario or for solving the problem that you are trying to solve then lastly you should be able to apply those methods you should be able to you know configure and optimize those techniques so once you have chosen a particular technique you should first of all the data literacy would help you understand the problem and the method that you have chosen to apply and then you will be able to find out or you know you'll be able to figure out the interplay between the two truth number two the paradox of experience now a big or major challenge with getting a job in data science is that most companies are looking for experienced data scientists who can actually drive decision making you know and refine their business models and help them you know provide more values to their users so how do newbies break into data science so you must have come across those memes that talk about needing the experience in order to get experience that kind of holds true in this particular case so how do we break into this field so the only solution is that you need to work or you need to produce a solid work portfolio that can actually showcase that you can do the job uh one very important quote that i came across the other day was you know in order to get hired you should start doing the job before you get the job truth number four programming is inevitable you need to learn how to program in order to become a good data scientist now up to a certain extent you can use tools like excel or business intelligence tools like tableau or power bi for your analysis or even modeling for some cases but it doesn't provide you the flexibility of you know running multiple analysis multiple models with different data sets and reproducibility has been a major concern so now we are looking at tools that can actually help you you know version data sets as well as um ml experiments and this is where uh programming comes into play and becomes required becomes mandatory in order to you know learn or build complete machine learning pipelines truth number five no matter where you come from domain knowledge is key so the last company that i worked at they used to send research papers to their data science candidates and by doing that they used to check the candidates ability to wrap their head around biomedical problems data sets and the theory which included a lot of biology and chemistry so this exercise used to inform them about the ability of the candidate to make decisions and you know drive growth of the data science team now business understanding here or domain understanding is basically the intersection of your data science paradigms and the practicalities of real world problems and having a good understanding of the domain basically helps you you know with a number of things first of all you will learn how to define the right questions you know choose the right data sources uh measure the right metrics prioritize the right product line and in turn it helps in business uh it helps in growing the business faster so a new uh you know a big chunk of newcomers basically find it really hard to specialize in one field and that's kind of the hard truth so it goes back to point number one which is you know become data literate and then try to narrow down try to specialize and find out which particular domain you want to apply your skills in so those were the five hard truths and if you like the content then you should consider subscribing to my channel as well as subscribing to my newsletter if you want all of this content everything interesting that's happening over a week delivered right to your inbox then you should definitely look at my newsletter i am sharing authentic data science content every week it could be illustrated tutorials it could be interesting reads that i come across or it could be some announcement around workshop webinars or any interesting course that i am conducting or i came across so yeah with that we have come to the end of this video and yeah i'll catch you guys in the next [Music] one