What Does It Take to be a Data Scientist or Analyst These Days?

What Does It Take to be a Data Scientist or Analyst These Days?

While there have been many new job titles emerging in the 21st Century (who could have imagined working as a “cloud analyst” even ten years ago?), none has been more critical to the aspirations of enterprises as that of data analyst or data scientist.

What special blend of skills do these roles require?  And is there a difference between a data “analyst” or data “scientist”? These are the questions addressed in the latest works from Brian Liou, Tristan Tao, and Elizabeth Lin, all with Leada.  In The Data Analytics Handbook for CEOs and Managers, the authors gathered to the views of prominent data experts.

The most important point coming from all these thought leaders is that data analysts and scientists are in the people business. The statistical and programming work is secondary, merely tools to support efforts to open people to new insights. “To some degree, analysis or data science is all about trust,” says Derek Steer, CEO and co-founder at Mode Analytics. “You can deliver the best insights, but if your customers don’t trust you, it’s useless. Analysis that does not turn into a decision is basically a waste of time.”

Establishing trust “is a tricky thing to do and requires a fair degree of interpersonal skill,” Steer continues. “It’s about communicating ideas clearly.”

In addition to being people-oriented, getting into data science or analysis requires a “freakonomics mindset,” says Dean Abbott, co-founder at Smarter Remarketer, Inc. “Love for data and an innate curiosity make for a great analyst.”

The thought leaders also discussed the difference between data analysts and data scientists. “Given a set of information like the order data for an e-commerce site, the data analyst knows how to answer important questions like, ‘how many products did we sell this year compared to last year’ or ‘which product categories have the highest profit margin?’” says Tom Wheeler, senior curriculum developer at Cloudera. “The data scientist, on the other hand, knows how to gather new data as well as query it. They approach the problem scientifically, designing experiments, forming hypotheses, and building systems to collect the data needed to validate the results. In other words, the data scientist knows how to ask questions as well as answer them, and has the software engineering skills needed to build new tools when needed. That’s a rare combination of skills, but one that’s increasingly in demand as businesses become more data driven.”

Rohan Deuskar, CEO and co-founder at Stylitics, defines the role of data analyst as “somebody who can sit down with a set of data, in some simple form like CSV, and is able to do two things. Firstly, understand the patterns and trends in the data. Secondly, be able to manipulate the data through various tools in order to understand the fundamentals of the dataset. The best data analysts act like detectives.”

A data scientist, Deuskar continues, “shares the same curiosity and intuition, but is much more technically proficient and has a very solid grounding in statistics. They should be able to say, ‘Yes on the surface this looks like a trend but for this reason it’s not.’ I would also look for someone who is comfortable with the latest machine learning methodologies and has applied those methods to real-world data problems.”

You don’t need a Ph.D to be a data a scientist, Wheeler adds. The strength of their background comes from “a mix of skills in statistics, software development, research, business, and so on. That’s a really rare combination and there are currently very few Ph.D programs that give you exposure to all of them.”

To develop skills, Abbott urges prospective data scientists and analysts to “take a few classes to get some fundamental understanding of data — not necessarily databases, but rather about data and what data means.” He also urges prospective scientists/analysts to “start building models. Work on projects. It helps to work with someone who has done it before.”

The ability to link these skills to business success are important. The most important skills “are understanding which factors help a business succeed and determining the key metrics by which those can be evaluated,” says Tom Wheeler, senior curriculum developer at Cloudera. These skills, he points out, “ultimately provide the foundation needed for constant experimentation and iterative improvements that will help uncover important patterns and help a business thrive.”

“If we don’t break the rules then the rules will break us,” says Ali Syed, founder and CEO at Persontyle. “There are too many mediocre and boring things in life to deal with and data science shouldn’t be one of them.” Data science “starts with creative imagination,” he continues. “Our ability to imagine and think beyond the obvious is one of our extraordinary powers as humans. One can learn tools and algorithms to deal with this data deluge and get hired based on skill-sets reflected on a resume, but you will only be able to add real and meaningful value based on your ability to improvise, adapt, and create.”

“The last thing we need is a breed of data scientists who are too skills-centric and have not dwelled much on the idea of being the biggest skeptic of data,” says Syed. “Before we end up with an army of tech-savvy data scientists who are nothing more than clever calculators, we need to ensure that data science remains a science and does not degrade into some gibberish practice that is more of a collection of techniques rather than a discipline.”