How To Start A Career In Data Science In 2022: Advice By 20 Experts
If you are a Data Science enthusiast, this is the time to get into this field.
The field of Data Science is with no doubt lucrative and has consistently ranked as the number 1 job and among the highest paying jobs in the world. Most companies are spending thousands of dollars to hire and up-skill people with analytics skills. Many people are also enthusiastic to get into this field.
If you are a Data Science enthusiast, this is the time to get into this field.
However, starting a career in this field is not as simple as the other traditional roles since the field is quite new. Many are making regrettable mistakes by taking the wrong path or having misunderstanding of the field with no guidance.
Despite the high demand in the field, many are still finding it difficult to land a job.
In this article, expert Data Scientists from top companies such as IBM, American Express, Fractal Analytics, Myntra, Forbes and top 10 Data Scientist share their experience to guide beginners who want to start a career or become a Data Scientist.
You can access the complete book here
EXPERT INSIGHTS:
QUESTION 1
Why Did You Choose A Career In Data Science?
There are a lot of lucrative careers out there and you could have become any of them so why Data Science in particular?
Response 1:
One of India’s Top 10 Data Scientists | Senior Vice President of Data Science @ Forbes, Nissan Motors and Mercedes.
You are right that there are a lot of interesting careers out there. When I started there was nothing like “Data Science”. I graduated in 2001 with Statistics as my major and Mathematics as my minor. I have been in this field for the past 21 years. During my time, Java programmers were the ones getting the highest paycheck and if you are not a java programmer then things become tough for you. Most people, including myself, were confined to Risk Management which was evolving at that time. There was nothing called Data Science or Analytics(although we do it unknowingly). Then with technological advancements, tools and definitions of Data Science evolved. Those of us who have a background in Statistics and Mathematics easily adapted to working on projects that involved analyzing data and then eventually we discovered the power of these tools and started paying attention to the various research papers that were evolving.
Deep Learning and all those fancy terms you hear today were not that exciting in those days and besides most companies were skeptical about its implementation. However, data analysis was still booming due to the direct and immediate results that we could achieve. So I started at HSBC as a Risk Management Practitioner. Then I further studied and 5 years after that I joined the Analytics team at Nissan Motors. That is where it began for me. My background is what helped me to develop much interest in this field and that is why I have been able to stay in this field up to now.
Response 2:
Machine Learning Engineer @ IBM
I started my career as a .NET developer. It was a good career but with time the data science buzzword started to come up and I decided to check it out. I was more intrigued when I got to work on one or two projects and saw how my results have direct and immediate impact. It is like when you try to dig deeper into data, data speaks by itself and that amazes me. Another reason being that the field is quite vast and I can explore any area as much as I want.
Response 3:
Data Scientist @ CIMB Lab
In my case, it was not straightforward whether I want to become a software engineer or something else and it was quite boring. I studied software engineering during my undergraduate degree, although I wanted to do something related to animals because I love dealing with animals but my dad said no, so I listened to him and did engineering. After my bachelor’s, I was lucky to join several South Asian companies. Initially, most of the work I got involved with was what we used to call “post mortem”: basically finding what happened in history. But later on, I got work with clients from the UK and most of the work was in predictive analytics, modeling, and dashboarding.
After that, I wanted to take it up fully and do my master’s but I did not want to quit my job either. So I started with an Executive 1.5 years program and afterward I joined CIMB Lab as a Lead Data Scientist. I joined the company in Malaysia, away from India and it was a bigger opportunity for me. I came back to India after a couple of years and joined 25 other Data Scientists in Bangalore to establish a similar A.I. lab. Since then I have been working as a Lead Data Scientist up till now. So that is how I found myself in this field.
Response 4:
Data Scientist @ American Express
Initially, I was in a dilemma to choose between Data Science and Software Engineering because my department at college was oriented towards software engineering and data science. When I started working on projects I became more inclined towards A.l. and my first was in Natural Language Processing(NLP) and after that, I also worked on a couple of Computer Vision projects which gingered me more. After that, I started exploring more online courses on Artificial Intelligence. I also discovered that it was linked with statistics which I very much enjoy doing and I could easily relate to it.
I, therefore, got a few internships in the Data Science field and things began to unfold. Henceforth I decided to stick to the field of Data Science.
Response 5:
Data Scientist @ Fractal Analytics
My interest is in technology and after my bachelor’s, I started building websites. I studied Physics and mainly Quantum Mechanics which is a fusion of mathematics and Physics. Most of the things I was doing in Quantum Mechanics were similar to what we call Machine Learning today but as of that time, we did not call it Machine Learning. In Quantum Mechanics we normally try to find approximate solutions to a problem and that is what is also done in Machine Learning, trying to find the most optimized path and approximate solutions to real-world problems. That is what is compelling to me in the field of Data Science and my background also makes it quite easier for me to continue in this field.
Response 6:
Data Scientist @ IBM, Myntra
When I finished my master’s in computer science back in 2013, there was nothing like Data Science, Machine Learning, or Artificial Intelligence. However, I was lucky to join the Data Information team at IBM Research Lab where I worked on Data Mining projects. We used to mine data from Twitter to fish out who and where there is likely going to be a protest and comments that are likely to cause social unrest. It was a very interesting project I must say. I did this for 2 years and I started my Startup. It was in the Adtech(advertisement and technology) domain. Here, my team and I used to work on over 20 terabytes of data every day. I learned everything from A-Z of Data Science. We did not have many resources to hire from outside so I have to do almost everything all by myself and sometimes with the help of my colleagues. We used to build models to predict whether a customer will click on a particular ad or not. I did that for approximately 3 years and then I joined Myntra which is part of the Walmart and Flipkart group as a Data Scientist. At Myntra, I am currently working on price optimization.
So this is how I found myself in Data Science and I am loving it.
QUESTION 2
What Skills Are The Most Valuable For Data Scientists?
If one wants to start a career as a Data Scientist, which skills should he/she focus on the most?
Response 1:
Benjamin Skrainka
Principal Data Scientist, Galvanize
Data scientists need expertise within multiple disciplines. You need to be good at databases. You need some knowledge of software engineering. You need to know some machine learning. And you need to know some statistics. “At the same time, I think curiosity is important. Data scientists are inquisitive. They are continuously exploring, asking questions, doing what-if analyses questioning existing assumptions and processes. They will always be learning and thinking about what new technologies are out there that will help them be efficient and help the business succeed. While there are a lot of great tools available, there is no substitute for thinking.
Response 2:
Cliff Click
Chief Technology Officer, Neurensic
Data scientists need a good blend of domain knowledge and a blend of business expertise. They need to be extremely inquisitive and relentless at figuring out how to solve a particular problem. That means digging into different approaches and alternatives — not just building models and running algorithms, but also interpreting the results to drive new business opportunities.
Response 3:
Jorge Castañón
Lead Data Scientist, IBM
Creativity is a key element of data science. You need to have the technical background, but you also need to be curious enough to explore at a deeper level. You have to be able to go to the next layer, go deeper and explore. A skilled data scientist explores and examines data from multiple disparate sources. They simply do not collect and report on data, but also look at it from many angles, determine what it means and then recommend ways to apply the findings.
Response 4:
Jonathan Dinu
Vice President of Academic Excellence, Galvanize
One of the key attributes that distinguishes today’s data scientist is strong business acumen, coupled with the ability to communicate findings to both business and IT leaders in a way that can influence how an organization approaches a business challenge. Data scientists often become the liaison between IT and C-level executives. Therefore, they need to be able to speak both languages and understand the hierarchy of data; they can’t just be the data expert.
Summary Of Skills To Master From Experts:
- Basics of Mathematics and Statistics.
- Strong Knowledge In Python, SQL, Databases and Excel.
- Master Data Visualisation.
- In Depth Knowledge of Machine Learning Algorithms.
- Pick An Area and Specialise (i.e. Natural Language Processing, Computer Vision, Big Data, etc.)
- Master the Art Of Storytelling.
- Master Communication and Presentation Skills.
- Master Teamwork and Collaboration.
- Develop domain knowledge as you progress.
QUESTION 3
What Resources Helped You?
When you began learning to get into the field of Data Science, which resources did you use. Any books, blogs, articles, courses or anything else that you can share?
Response 1:
Data Scientist @ American Express
I think the most important thing to do when you start is to get your statistics and coding skills in order. By that I mean make sure you develop a solid background in maths and stats alongside coding skills. For me I used random Youtube videos to learn coding in Python. I already have the basics of statistics so I need not to worry much about statistics and mathematics but I still read a couple of statistics books such as Practical Statistics For Data Scientists by Peter Bruce and Andrew Bruce. As for Python, you can easily get started with Youtube videos or Udemy courses such as the Python Crash Course
For Machine Learning, I used Andrew Ng’s course on Coursera which is a great resource to get started with Machine Learning.
Response 2:
Data Scientist @ Fractal Analytics
I think the best resource is getting the hands-on experience right from day one. In my case, I learned from my mentors who were active Data Scientists. I did not do any certifications and up to now, I do not have a single certificate in Data Science. I focused on working on real-world problems rather than choosing hypothetical datasets. I used to choose a particular problem around my environment and try to collect the data by myself and then prepare it and build models to make some predictions. That gave me a good sense of how real-world data science is and this made it easier for me to get into the field.
I used to participate in Hackathons, some of them from Hackerank, Kaggle, etc. I also used to read blogs a lot from Medium. I used to make sure I googled as much as I could on a particular topic. For instance, if I pick say Logistics Regression, I will google lots of resources on that and learn into details, especially the mathematics behind it.
For Deep Learning, I used the Deep Learning ebook by Ian Goodfellow and Yoshua Bengio. I joined a startup and started working on projects related to deep learning, particularly, computer vision projects. So that is how I did it in my case.
Response 3:
Data Scientist @ IBM, Myntra
I was more into hackathons so I used to follow Kaggle a lot. I also used to read lots of blogs from Medium, Analyticsvidya, and KDNuggets. These resources helped me to understand some basic concepts and how people are also implementing concepts that I was not familiar with. I particularly did not take any Data Science courses since they were not available at the time that I started in 2013. But as I mentioned earlier, I was lucky to join the Data Information team at IBM Research Lab which got me started on fertile grounds.
Response 4:
Data Scientist @ CIMB Lab
I did had a coding background so I did not have difficulties when it comes to coding. However, in my time we did not know which programming language is good so I started with Java then I moved to R programming then later to Python. But now it is clear that Python is the best when it comes to Data Science. So I will say start with Python which is also easier to learn compared to all the other programming languages so far. For Python I suggest just start with the Python documentation or if you are a complete beginner take some Udemy or Coursera courses. I also used to read a lot about the details of the various algorithms, for instance what is the math behind a particular algorithm and why is one algorithm better than the other and in what specific ways. I also made sure that I have enough understanding of some key maths and stats concepts such as linear algebra and calculus, that helps especially when you are trying to minimise cost function in machine learning then calculus comes in handy. I think algorithms are not a problem, that comes with experience, as you work on more and more projects, you turn to know which algorithm to use for what problem.
Response 5:
Machine Learning Engineer @ IBM
I mainly took some Data Science courses from Coursera and Udemy. For practicing, I used Kaggle. I studied what people were doing and started doing my own thing with some of the datasets and competitions over there. I also read a lot of good blogs from Medium to understand certain topics. These days most of the concepts are on Google and a simple search on a particular topic can help a lot.
QUESTION 4
What Are Some Of The Challenges You Face On A Day-To-Day Basis?
Response 1:
Michael Schmidt
Data Scientist/Founder, Nutonian
One of the biggest challenges as a data scientist is applying the domain expertise to solve a problem. We have a plethora of algorithms and techniques to get value from data, but we need solutions to help us apply that to applications — to connect the dots from the statistics to the business opportunity. “Solving problems and predicting outcomes using sophisticated models require both an understanding of the capabilities, tools and techniques behind data science and the ability to get out from behind the keyboard and ask questions to inform the data process. Interpreting the problem is as much an art as it is a science.
Response 2:
Andy Gants
Principal Data Scientist, Spare5
One of the larger challenges I faced in my current job was that the probability and statistical estimation tools that I used previously in my earth science research were the same tools, but they do not necessarily perform the same way on these new problems that involve user and answer estimation, quality estimation in these intelligent crowdsourcing problems. So the tools are the same, but the application of those tools is different. Also, I didn’t have a software development department that I was iterating with in terms of implementing analyses into specific software features. Learning how to perform with the software development department proved to be quite a challenge — but a fun one.
Response 3:
Roman Schindlauer
Program Manager, Dato
One of the biggest obstacles to analytical productivity is refining and formatting the data required for high-quality analytics. The lack of a universal or standardized programming language specifically geared to the data science domain doesn’t make it any easier. Even with the best tools today, there is really not a good way around cleaning up the data manually. It’s a continuous cycle of collecting and cleaning data and trying to figure out if it will yield significantly relevant insights. Or will you need to go back and change the parameters or massage the data more? I think we’re getting to a point where the tooling support is helping with that, but it still requires a great deal of manual manipulation.
QUESTION 5
How Do One Crack Data Science Interviews?
Cracking Data Science interviews seems to be a concern of many starters. How should one approach a Data Science interview?
Response 1:
Data Scientist @ IBM, Myntra
I think what makes Data Science interviews a bit tricky is because it varies very much from company to company. Unlike Software engineering where you can say that you are likely to be asked about Data Structures, algorithms, and coding skills, in Data Science, every company has its requirement and certain skills that they are looking for in a candidate and that makes it very undefined.
I will say that first take time to research about the company. If the company is more into scale, they will likely ask you about your experience in handling large datasets. If the company is more into consulting and service-oriented, it is likely they will be interested in skills such as SQL, Excel, some coding, and communication skills. If the company is domain sensitive like healthcare and finance, then aside from your coding and interpersonal skills, they are likely to ask about domain knowledge.
If you take time to do this basic research, you will get to know what to expect during the interview rounds and it will not take you by surprise.You can also search for some companies when you are starting your career and develop skills that match their Data Science requirements. In that way, you can easily match their job requirements and likely get hired.
Response 2:
Data Scientist @ CIMB Lab
What will help you during the day of your Data Science interview is if you have some good projects to talk about. It is important to have a Github repository that you have stuffed with some good projects. If you have projects to talk about already, it will help you to drag the conversation to your comfort zone. The interviewer might have a different experience from what you have and if you let them lead the conversation, they will be asking you things that you might not be familiar with and that will be a recipe for disaster. Your agenda should be to bring them to what you know so that you can talk about it in detail.
Response 3:
Data Scientist @ American Express
During Data Science interviews, you will likely have a technical round and a CV/Resume round. In the technical round, I will say master the algorithms very well and not just how to import and use them to build models but the mathematics behind them is very important. You will need to explain why you are using one algorithm instead of the other. What hyperparameter optimization you have considered. Lastly, how you can use storytelling to explain your output is very important.
For the CV/Resume round, it all depends on the kind of projects you have done already and how well you can explain them for the interviewer to believe that you did it by yourself and you really understand what you have done. If you do not have any projects in your GitHub or portfolio, you are opening up for disaster. The interviewer will bombard you with questions that you might not have any idea about.
There are many more to discuss when it comes to starting a career in Data Science and Analytics.
The insights shared by the experts are have been put together as a book and a video and the full version of it can be accesses below:
If you like this article, kindly give it a thumps up. Thanks in advance.
If you wish to write in this Newsletter, kindly reach out to us via Whatsapp: +919467891831
You may want to check: Full Stack Data Scientist A-Z™ BootCamp
