Becoming a data scientist: The career path for job changers
While Willie Sutton may never have actually said he robbed banks “because that’s where the money is,” the sentiment is a clear one: Find the most profitable source of personal opportunity.
If your IT career has stalled, you might be tempted to shift to a more remunerative path. Nowadays, there’s plenty of opportunity in data science. Here’s what it takes to make the transition.
Do the numbers
Anytime you make a career change, it should be in the direction of a growing field—more jobs available—accompanied by more interesting work, meaningful results, and ideally higher pay. That’s certainly the case for data science jobs.
Data science job openings are expanding faster than the number of technologists looking for them, IEEE Spectrum reports. Citing research from job search site Indeed, it notes, “Job postings for data scientists as a share of all postings were up 29 percent in December 2018 compared with December 2017, while searches were only up around 14 percent.”
The companies aren’t finding those specialists. Many are facing shortages of people with data science skills: “Data science skills shortages are present in almost every large U.S. city,” according to the 2018 LinkedIn Workforce Report. “Nationally, we have a shortage of 151,717 people with data science skills, with particularly acute shortages in New York City (34,032 people), the San Francisco Bay Area (31,798 people), and Los Angeles (12,251 people).” And the need has only become more urgent: The LinkedIn Most Promising Jobs of 2019 lists data scientist at the top, with a median base salary of $130,000.
What do you mean by “data science”?
While data science is undoubtedly an appealing career, there are all sorts of subspecialties.
“If someone came to me and asked for advice, my first question would be, ‘What do you think data science is, and what would your ideal job look like?’” says Carly, a data scientist. “There's still no commonly understood definition of data science, and the role differs greatly from one organization to another. Also, public perception of what a data scientist does is oftentimes very different from reality.” The role might be oriented around machine learning, or product analytics, or traditional statistics, all of which are really different, Carly adds.
If you aren’t sure how the field breaks down, don’t assume it’s your own ignorance. Even the professionals dispute what is and isn’t in their realm. However, according to one professional cited in a Harvard Business Review article titled "What Data Scientists Really Do, According to 35 Data Scientists," data science can be categorized in three ways: “business intelligence, which is essentially about ‘taking data that the company has and getting it in front of the right people’ in the form of dashboards, reports, and emails; decision science, which is about ‘taking data and using it to help a company make a decision’; and machine learning, which is about ‘how can we take data science models and put them continuously into production.’”
Which of those appeals to you most?
For example, data scientist Bailey Karl emphasizes the value of data organization. “There are lots of steps involved in data preparation, from translating certain system codes for making data usable to dealing with erroneous or incomplete data,” Karl says. That means an ongoing demand for skilled individuals who can eliminate bad data and organize it so valuable insights can be extracted.
Some practices apply to any IT pro who wants to move into a new area. Read the “expected” books, for instance. Also, points out a librarian friend, check for training classes online, particularly on Lynda.com (now LinkedIn Learning), because the latter often is free with a public library card. Among the online courses specific to this subject are MIT’s Matrix Methods in Data Analysis, Signal Processing, and Machine Learning. There are also quite a few data science boot camps.
Alternatively, says career counselor Ebony Johnson, “Search on LinkedIn or within your social and professional networks and contact professionals in that field. Ask to meet and conduct an informational interview to learn what it takes to land the job.”
Look for data science meet-ups in your area, or attend a data and analytics conference. In addition to learning directly from practitioners, it’s an opportunity to befriend specialists, which is good for personal networking. That is, your new friends may be so impressed with your enthusiasm to learn the subject that they’ll ask for a copy of your resume and press it into the hands of hiring managers. (This is not apocryphal advice. I got my very first professional programming job that way.)
Leverage your existing knowledge. For example, someone who’s bored with banking IT likely has database experience as well as financial tech domain knowledge. As one experienced data scientist suggests, “If you have some knowledge of data analysis and statistics from your studies, you might be able to move to an analyst role of some kind where you work with the same databases and in the same industry. With experience with databases, you might also be able to grow into this direction—think database administration or data engineering.”
Another (more introverted) way to go about this is to reverse engineer the required skills. Helena Goote, a quantitative research analyst working on healthcare analytics, recommends that you look at job descriptions, particularly those for which you’re not (yet) qualified. When you see a job you’d love, look at what the job requisition asks for and try to match those skills.
So, which skills?
Because the term "data science" lacks a standard definition, it’s hard to provide a shopping list of skills everyone needs to learn and the technology expertise to acquire before looking for a job.
However, some baseline knowledge across data and analytics is required. It’s likely you have at least some of this knowledge if you’re coming to data science from another area of IT. For example, says Carly, “SQL fluency is a requirement for most data science roles now.”
So is math—at least for the more tech-heavy jobs. “Data science is glorified statistics with a good dose of linear algebra, and with R or Python code instead of SPSS (or similar),” says Isac Artzi, associate professor of computer science at Grand Canyon University. “Therefore, someone with programming experience must first learn foundational statistics, primarily linear regression. This should be followed by foundational linear algebra, focusing on vectors and matrices. Machine learning and neural networks are just fancy words for linear regression and matrix algebra.”
“Learn about machine learning, neural networks, and big data software,” adds Goote. “Learn Python and R.”
Python and R are recommended frequently. Fortunately, Python is among the most popular languages these days, so you likely know it already. “Most contemporary data science uses Python and, in particular, Jupyter Notebooks to test and analyze data, so you need to learn how to use both effectively,” says an analyst named Zach. “A good foundation in NumPy, SciPy, Matplotlib, and Scikit-learn is super-useful as well.”
Zach has specific recommendations to come up to speed and get a solid foundation, starting with Hastie, Tibshirani, and Friedman's "The Elements of Statistical Learning." For practical projects, he suggests Géron's "Hands-On Machine Learning with Scikit-Learn and TensorFlow."
“The first seven chapters of that book are pretty useful for learning some useful, applicative regression analysis,” he says.
For your own self-teaching projects, it might help for you to be a baseball fan. “There’s a lot of pop-lit around sabermetrics that is a good example of how data can be applied to a specific domain—in that case, baseball,” says Carly. “Product data science is kind of like that, except for business.”
Ready for a job change?
One entry point for career changers may be smaller companies. They are anxious for data analysis and may provide an opportunity to use a variety of data science tools. “Entry-level data scientists are getting more preference at smaller firms and startups mainly because data scientists come with a broad range of skills that seem to be advantageous for smaller organizations,” says Karl. “The hiring processes for smaller companies are also relatively faster compared to larger organizations.”
Alternatively, it may make sense for you to make the career shift in stages. “If you’re actually starting from scratch, it may not be the best idea to aim for a full-on data science role,” says Karl. “Instead, you should target relatively easier positions in the field, like data analytics or data visualization experts, among others. These positions usually involve working alongside data scientists, which can greatly help you in gaining some experience before you aim for the bigger goal.”
Moving into data science: Lessons for leaders
- Reverse engineer job listings. Find your dream job, identify the skills you lack to qualify for it, and then learn those things.
- Build on your existing knowledge. Your business domain skills or other tech knowledge can inform your new data science job.
- It may make sense to apply for jobs in smaller companies rather than large ones. The smaller firms are less stringent about job requirements, and you have a better chance of learning a wide range of data science skills.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.