t's the most popular job for 2016, per Glassdoor.

And, according to Eric Haller, a vice president at the global information services company Experian who runs their DataLabs unit, it's also one of the hardest positions to fill. Recently, I asked him about the role and why it is so important that many startups will likely need to hire one in 2016.

Q. What is a data scientist?

What are the typical job duties?

A. I'd say a data scientist is an explorer, scientist and analyst all combined into one role. They have the curiosity and passion of an explorer for jumping into new problems, new data sets and new technologies. They love going where “no man has gone before” in taking a new approach to taking on age old challenges or coming up with an approach for a very new problem where nobody has tried to solve it in the past.

They have the discipline and knowledge of a scientist. They are comfortable in the rigor of forming a solid hypothesis, creating a valid credible path to a solution and being able to objectively assess performance of what they've built.

They can write their own code and develop their own algorithms. They can keep up with the scientific breakthrough of the day and regularly apply them to their own work. And as an analyst, they have a penchant for detail, continually diving deeper to find answers. Finding treasure in the data, analysis and the details give them an adrenaline rush.

Our data scientists tend to operate with a noble purpose of trying to do good things for people, businesses and society with data.

Q. Why is it an important role?

A. Our world has become increasingly digital. We shop and buy more of our products online or via mobile, we decide where to eat, buy a car, hire a plumber by consulting what's available to us digitally and any critical decision that is made — a credit decision, admissions into a hospital, accessing your social security benefits statement — are all decisions made electronically.

Data fuels more and more of what drives the economy. I'd love to challenge someone to try to live a life without leveraging data. Pay with cash, unplug from everything digital — the way we live today that would be an enormous challenge.

As a result, making decisions around risk, financial transactions, detecting fraud, marketing to consumers and businesses, improving the customer experience — can all be influenced, shaped and driven by our ability learn from data and optimize the actions that are being taken.

Q. Who needs one?

A. It would almost be easier to ask who doesn't need one. In the United States, almost every business has a digital footprint. Even if they don't conduct their business over the Web or via mobile, they likely try to build awareness that they exist through those channels. And if they aren't consumer facing or don't feel the need to market themselves, they may likely benefit from analyzing the data captured around their supply chain and sales funnel.

If anything, the data science proliferation we are seeing that has deep roots in Silicon Valley is only growing broader and wider over time.

At Experian, we use data science and the insights discovered from it to help a consumer secure an affordable loan, improve their credit score or protect their identity; or for a business to mitigate risk, help prevent fraudulent transactions or even to ensure they are marketing their products and services to the right consumers at the right time and across the right channels.

Q. Why would a startup need to hire one soon?

A. It depends on the start-up and what they do. Start-ups related to cloud-based or mobile offerings likely need one or more data scientists right out of the gate. Best to assess up front how data is being captured, time period requirements for storage, attributes generated and end performance benchmarks for fraud/risk management and the customer experience. Also creating a computational environment for the data scientist to have a safe zone for them to conduct their work so they are up and running even as version “.5” is being created in advance of a “1.0” launch. It's even more likely that start-up 1.0 will use what comes from the data scientist in start-up 1.1 and every version that evolves from that point.