We’ve seen all the stories: Data scientists are among the best, most in-demand jobs in America. With companies offering high salaries and education institutions scrambling to create programs, everyone is looking for individuals who can glean insights from the troves of data that organizations are collecting today. But the title “data scientist” may have become too all-encompassing for one person to fill, some say.
“I’ve never been a fan of the data scientist title,” said Meta S. Brown, business analytics consultant and author of Data Mining for Dummies. While it makes sense for managers who need a separate title for a particular job description to work with HR, the constant hype around the title today has created some of the problems companies have filling positions, she added.
“We’ve pushed the idea that data scientists are something different from any kind of analyst who ever existed before, and are somehow magical and better,” Brown said. “It has lead to a lot of people wanting to be this magical better thing, and a lot of organizations imagining that there’s something that is unrealistic that can solve all their problems.”
SEE: Job description: Data scientist (Tech Pro Research)
“The data scientist title is certainly a sexy one,” said Kristen Sosulski, clinical associate professor of information, operations, and management sciences in the Leonard N. Stern School of Business at New York University, and author of Data Visualization Made Simple. “However, it is a general title that can encompass a variety of responsibilities and require a diverse set of skills.”
Other titles that might be more descriptive of the actual work involved in data science professions include data analyst, business analyst, data engineer, or analytics lead, Sosulski said. More senior titles include head of analytics, head of data, and vice president of analytics.
How to write a data science job description
Part of the problem is that managers often write data-related job descriptions without fully thinking through the needs of their organization, Brown said.
“Very often, these jobs that no one can fill are signs of either imagination or desperation or a little of each,” Brown said. Recruiters have come to Brown looking to fill a data science position, she said, “and they send me a requirements list that range from reasonable to completely imaginary.”
Brown said that when she suggests capable candidates to them, they very frequently say they want “something more.” But when asked what that means, they often say they do not know, she added.
SEE: Big data policy (Tech Pro Research)
“The truth is there’s a great deal of very capable analytic talent out there, so I don’t believe that there is a lack of talent,” Brown said. Training employees on data skills that will be needed in the future is often a better option than seeking employees from abroad or other solutions companies have turned to, she added.
Hiring managers should be realistic about the type of skills really needed for the position, Sosulski said.
For example, if you want your data scientist giving presentations, then strong information visualization and communication skills are necessary, she added. If you are hiring only one data scientist, then you may want to consider a more senior position with skills in programming, machine learning, data modeling, data architecture, and statistics.
SEE: Prescriptive analytics: A cheat sheet (TechRepublic)
If you are filling one of many data scientist positions on a team, you may consider focusing on depth more than breadth, and recruit for the one or two skills you need that would benefit the whole team. That way, both the new recruit and existing team could learn from each other and broaden their skills, Sosulski said.
To learn more about the skills needed to become a data scientist, check out this TechRepublic cheat sheet.