A short and sweet guide covering answers to 15 of the most common questions about data science
Getting a job in data science is hard.
Not only do you have to master a difficult mix of programming, mathematics, data analysis, and graphic design, but you also have to compete against hundreds of candidates with similar qualifications, abilities, and reasons for being there.
With such an interest in this exploding tech field, hundreds of free resources are available online to provide answers to all of your questions about data science. However, are any of them providing definitive answers that have been checked by others in the industry for accuracy?
With hundreds of resources to sift through, it can be tough to get a short, simple answer to a question. That’s where this guide comes in — to provide you with the basic information you need to answer all your burning questions about getting a job in data science.
To get a job in data science you need related experience, whether that’s employment in the field, a related degree, or an impressive portfolio containing projects that speak to the hiring manager.
Most companies need their data scientists to be able to hit the ground running, which is why many require new hires to have some years of experience. Therefore, it can be beneficial to take jobs at the beginning that get your feet wet while giving you real-world experience that can be applied.
Doing pro-bono work, adding your degree capstone project to your portfolio, or working as a data analyst or data engineer for a couple of years can help you get your foot in the door.
Experience is how you get a job in data science. Creativity is how you gain experience.
Gaining related experience can come in many forms, including networking, contributing to significant blogs or publications, building your own consulting business, working on projects to showcase your talents, interning, volunteering, or doing pro-bono work.
Many hiring managers are becoming used to seeing candidates that have non-traditional backgrounds, which means that alternative experience isn’t going to count against you. The key is to show that your relevant experience is what will help you make an instant impact on the company.
Any degree is a good degree for data science as long as you can fulfill the requirements of a data science position.
Did your degree teach you linear algebra, calculus, and statistics?
If so, then you’re probably set for a career in data science.
If not, then can you teach yourself the skills needed?
Unrelated degrees won’t hold you back from getting a job as a data scientist as long as you can carry out the daily tasks required on the job. You could have a degree in underwater basket weaving and as long as you can test hypotheses, write some code, and tell a story using data, you’ll be able to get a job in data science.
Most data scientists you talk to won’t have started as data scientists. More likely than not, you’ll be talking to former engineers, scientists, policy writers, teachers, and more.
Again, as mentioned above, as long as you can fulfill the job requirements of the position you’re applying for, then entering the field of data science from something completely unrelated won’t be an issue.
Luckily for you, there are hundreds of resources online that can help you make a transition with no prior data science experience.
A Master’s in data science from a relatively well-known school can help solidify your position as a strong candidate for any data science positions you apply to.
Additionally, a Master’s can increase your earning potential right from the very beginning of your job search.
Bootcamps tend to be hit or miss on the benefits they bring you as a candidate, with most in the field tending to agree that bootcamps are a waste of time and money. The same goes for online certificates which are often held by hundreds of people vying for one entry-level data science role.
Bootcamps and online certificates aren’t what is going to make you stand out against all the other candidates who believe they deserve the position.
If you’re considering doing a data science bootcamp, the real question you should be asking yourself is what value you’re hoping it will add to your job hunt. Most bootcamps promise placement after completion, though few studies are available that prove that this promise benefits its students.
Therefore, the better question to ask is: What skills am I missing that will help me do better in data science interviews?
You’re better off self-studying the topics you’re lacking in and then getting some hands-on experience instead of paying thousands of dollars to be taught by former students of the bootcamp program who can’t manage to get a job either.
There’s no easy way to say this — you need to know most math that you think you need to know in order to become a data scientist.
No shortcuts genuinely exist when it comes to learning the math required for the job.
The amount of math required depends on the field you enter, the seniority of your job, and the general job duties. Some positions will only require basic calculus, algebra, and statistics, whereas others will require even advanced levels beyond that. It all depends on what daily duties are expected of you.
If you’re afraid of math, aim for data analyst positions which generally require only basic mathematics. If you’re up for a challenge, aim for positions where machine learning and artificial intelligence will be your bread and butter.
The short answer? Get good.
The long answer? As mentioned above, the amount of math you need to know to become a data scientist differs from position to position. Therefore, you should be able to tailor your learning to the amount of math required for your desired job.
However, the hidden answer if you truly believe that you’re bad at math is to focus more so on how things work and less on why things work—understanding the how versus the why of math are two very different things that aren’t necessarily required for each data science position. If you can teach yourself different mathematical models, how they work, and the results they should produce, you’re likely golden. It’s only when you get into the truly theoretical parts of data science that you’ll need to know why something works.
In short, math is all about practice and if you want something badly enough, it won’t be a burden to practice until you can successfully do what is required of you.
Python, R, and SQL.
All data science job requirements will be different but these are the three reoccurring culprits.
The answer to this is subjective based on which field you want to enter.
If you want to be a general data scientist who will be able to apply for most jobs on the market, Python is a safe bet. If you’re looking to enter science-related fields, R will be most commonly used.
Many data scientists will suggest that you learn both but that can feel overwhelming when you’re first starting out. Therefore, the best idea is to pick one, learn it inside and out, and then decide where your career will take you before learning the other (if necessary).
Yes, and link to it in your resume and cover letter so hiring managers can see what you’ve got.
The key to an intriguing portfolio is to include unique projects that aren’t using the same common data sets you see on Kaggle. That means nothing to do with the Titanic, stock markets, or facial recognition.
Hiring managers are tired of seeing the same old projects done just slightly differently from candidate to candidate.
Therefore, be unique, stand out, and make your portfolio memorable by thinking outside the box and developing projects that will truly speak to the hiring managers of your chosen industry.
Anything that interests you!
As mentioned above, for greater success with the hiring managers, avoid anything concerning the Titanic, stock markets, or facial recognition.
Hiring managers can tell right away if a project is something you truly put some thought into or copied and pasted the code from someone else’s code repository. Therefore, pick an issue, find some relevant data, and build a project that not only intrigues you but is also unique and offers hiring managers a fresh take on a problem that is relevant to their industry.
The ability to write code quickly and correctly will add to your impact as a data scientist.
While no-code data science is beginning to take root, the ability to produce unique code for data analyses is still a major mainstay in the field.
Luckily, learning to code is arguably the easy part of the whole process, and with hundreds of free resources only, learning to code will be the least of your worry.
Data scientists are an interesting hybrid between boardroom executives and underground software engineers — they need to be able to work within a team and present their findings to stakeholders, but they also need to be able to buckle down and work independently to complete projects.
Therefore, not only do you need to be good with people, but you also need to know how to communicate. Without proper communication skills, stakeholders won’t know what the data is telling them or which critical decisions they should be making.
So, if you haven’t already, it’s time to brush up on your teamwork and communication skills before you hit the data science job market.
Every company has its own special recipe for conducting data science interviews, which means that you need to be prepared for a variety of questions, scenarios, and interview formats.
The key is to focus on building mental models, practicing algorithms, and getting used to real-world problem-solving. Memorizing data science interview questions is useless because hiring managers aren’t looking for you to regurgitate some rote memorization — instead, they’re looking for you to reason, analyze, interpret, and collaborate to solve a problem.
By practicing these cornerstone skills, as well as learning how to communicate your thought process, you’ll be able to master most data science interviews you come across.