#DataHired: The Need to Know Basics

At the turn of this year, the Data Science Degree Programs Guide posted an update saying that by 2026, employment for information scientists will have risen by 19 percent, accounting for 5400 new jobs by the end of the decade. These projections may change in the light of new unemployment statistics estimated by the International Labour Organization; they predict that the coronavirus will engineer the loss of approximately 24.7 million jobs worldwide (cnbc.com). However, such uncertainty has yet to stifle the enthusiasm of those curious to learn about the newest data science tools, which inspired my #DataEveryone chat last week about navigating the industry’s job market.

Given that the top 10 percent of current data scientists in the field make salaries over 120,000 a year, it’s no surprise that talent across industries clamors to claim a spot. But funny enough, one of the major barriers to entry stems from confusion when interpreting the job postings themselves. What differentiates a data analyst from a machine learning specialist? And after determining which type of position might suit you, how will you ensure you’re prepared for the tasks involved? Experts from the #DataEveryone community have answers.

What Makes Applying for Data Science Jobs So Hard?

The confusing plethora of titles certainly doesn’t help. But the source of the issue comes from a deeper lack of consensus across the growing industry. “I think it is so difficult because of the variability within the roles, departments and companies,” @pip_alise commented. “The skills of a data scientist are truly on a spectrum, which can be hard to identify and best align with a specific role based on a job description.”

@DataSciBae agrees, adding, “There are so few hard rules for what each job does. So many startups want to use cutting edge AI, but have so few resources that folks typically do so much more than just analysis or modeling.”

@thedariaedits offered up some advice for making sense of the job postings: “First you have to understand what type of position they are describing,” she said. “Will you be working within a team training in languages you’re not familiar with, or will you be one of two data people tasked with doing everything?”

How Can Companies Improve the Application Process?

Let’s first start with the problem, described perfectly by @DataSciBae: “One of the biggest tragedies of trying to get a job in data science is that companies don’t even know what they need,” she described. “I’ve been the wrong experience fit in so many interviews, getting to the last stage just to hear ‘We’re looking for someone with deep NLP experience,’ which I clearly don’t have.”

How can we mitigate this disconnect? By taking the extra time to seek out experts who know better. “I really think people need to consult active data scientists before they start a job search for candidates,” said @thedariaedits. “It would clarify so many points of confusion if experts walked through the postings beforehand.”

Furthermore, more effort from top management may be required until a standard for data science roles is established. “Typically it is the non-technical members of HR that write the job descriptions or someone not involved with the intricacies of the role,” explained @pip_alise. “I’ve seen people copy and paste chunks from other job descriptions, which may not be relevant.”

When companies, such as Royal Caribbean in Miami, hire for multiple data roles at once, using separate job postings to distinguish between them, they are setting the foundation for much needed clarity. “Companies should outline what tasks and projects they should contribute to and create requirements based on that,” said @DataSciBae.

When is a Company Ready to Onboard a Data Person?

I started this blog from a job searcher’s perspective, but if you’re in charge of your company’s hiring, I have gems for you too.

“I think they are ready once they know where their data will come from and how to evaluate its quality,” said @ogustavo_com.” @pip_alise added, “They need to understand their own goals first, and qualify the quality and accessibility of their data.” Moreover, she believes that the various data roles should be onboarded in a logical succession. “A lot of companies don’t invest properly in data engineers and data analysts first before jumping to data science to solve all their problems,” she explained. “Personally, I see data science as necessary for optimization, automation and deeper learning. It’s essential for maximization, but for that to happen, you need to have a solid base to build on, which is where the engineers and analysts are pivotal.”

@DataSciBae has experienced benefits and challenges both of being early to arrive at a company and on the flip side, integrating later. “I thought being too early had some advantages, like being able to plan and collect data well from the get go. When the data is too disorganized, it’s hard to come in as a newbie, but it poses a good opportunity for someone more experienced.” @ogustavo_com believes that the challenges described by @DataSciBae could be applicable to the actual hiring process. “Before working for a think tank, they gave me a sample dataset and asked me to clean it, analyze it and run a proper model,” he described. “I always thought this was a good way of selecting people.”

So What’s Our Consensus?

We, just like the data science industry at large, are still working on it. But here are some final takeaways from the #DataEveryone chat that could prove useful:

“Data engineering allows for data access and maintenance, data analysis determines the “north star” metrics and day-to-day health of the company and data science involves long-term optimization and iteration.” - @pip_alise

In my enterprise environment, there is a lot of architecture that supports the data and not all data scientists know that architecture or want to deal with it. Data engineers can get it all flowing. Nothing is worse than hiring a data scientist with no data ready to do science.” - @jeremy_data

"Unicorns don’t exist. I’m a generalist, which is as close as you’ll get. It only works because I focused on startups for 10+ years. I can do most things to MVP level, but am not nearly as refined as focus expert.” - @Randy_Au


Looking for more tailored advice?

I’d love to hook you up with some one on one #DataHired tutoring. What’s in it for you? 30-mins or an hour of job-search consulting based specifically on your background, current skills and hiring needs. Desperate times don’t need desperate measures: I’m offering a full $20 off all sessions with the code ASKANYTHING.