The data science world is a strange place. No one is entirely sure what data science means or what data scientists should do. People come to the field with different skills and a wide variety of experiences. In fact, the field requires such a broad array of skills that it’s better suited to teams than individuals. While the insights delivered by data science are vital to modern business, the field itself has yet to settle into a defined shape.

To learn more about the data science field and the people inside it, Dave Vellante (@dvellante) and James Kobielus, host and guest host of theCUBE, from the SiliconANGLE Media team, visited the BigDataNYC 2016 conference in New York. There, they met with a panel of experts from the data science world to look behind the data science curtain.

An evolution in data science

The roundtable discussion opened up with a look at what the panelists felt were major changes in data science. Gregory Piatetsky-Shapiro, president and editor of KDnuggets, described data science as the second-oldest profession. People have a built-in need to find patterns and organize data. People do this well on the small scale, but not the large scale. He felt that was the challenge of data science.

Craig Brown, author of Untapped Potential — The Supreme Partnership of Self, mentioned that data science isn’t just about understanding and dissecting data, but now it’s concerned with helping people see more in the data than just raw information.

Joe Caserta, president of Caserta Concepts, saw a big shift in how they were preparing data for machines now, instead of other people.

A growing skill set among data scientists

The panelists discussed the transition of business from being driven by conventional wisdom to the new data-driven style. Along with that change, the needs of data science have expanded. A broad skillset is required for what is becoming an interdisciplinary position. Brown explained that talking about data science means talking about a collection of experiences. There is a wide combination of skills, enough so that one needs that combination among a team.

The group briefly explored their general skillsets, being very adept in mathematics. Jennifer Shin, founder and chief data scientist at 8 Path Solutions, LLC, related that as data science became a real career, she took a different approach. She learned to write applications. This helps her because a person can’t do the same kind of data science while developing a product; the wide range of models and flexibility isn’t there. By doing the development herself, she saves the team from that part of the process.

Miriam Friedel, research scientist at Elder Research, backed up this concept. She stated that a company needs someone who can deploy the applications built around their data models. A business needs a team of people to take advantage of different skillsets. Friedel also pointed out that once a model application goes into production, there has to be a model-management process to adjust the model for external factors and changes over time.

Delivering data insights

Data science is more than just numbers, it’s also about communication. The data science team must be able to present their findings to the company leadership. Caserta felt there is a science behind data visualization. The team has to consider their audience. Who is the consumer of that data? They have to tailor the information to that person.

On the other side of the communication coin is a notion of the “citizen data scientist,” a hobbyist researcher who engages in data science. The panel roundly dismissed the idea. They felt data was important, too important to have someone without proper training interpreting it.

“We don’t ask citizen-dentists to clean our teeth or citizen-pilots to fly our planes,” Piatetsky-Shapiro said.

Friedel agreed, saying it was too easy to make a spurious correlation and ascribe significance to it.

Finally, the panelists considered their final thoughts on the discussion. Friedel mentioned that it was interesting to hear her fellows talk about the same themes. Although everyone came from different places, they were experiencing the same challenges.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of BigDataNYC 2016.

Photo by SiliconANGLE