Think Like a Data Scientist (Because that’s Who Will Determine the Winners from Losers in the Data Economy)

Think Like a Data Scientist

There’s been plenty of buzz about data science and data scientists in recent times. While images of brainiacs in white coats pouring over rows of statistics come to mind, there’s much more to it than simply having a few smart people toiling in the dark recesses of some corporate lab. Data science promises to sharpen – in very deep and significant ways — the way we do business. For starters, data science encourages experimentation, in the form of deductive, or hypothesis-based reasoning to approach business problems. As problems reveal themselves to experimenters, inductive or pattern-based reasoning comes into play.

That’s the message in The Second Edition of The Field Guide to Data Science, just released by Booz Allen Hamilton. “Data science is an auspicious and profound way of applying our curiosity and technical tradecraft to solve humanity’s toughest challenges,” the book’s authors state. “The growing power, importance, and responsibility of applying data science methodologies to these challenges is unimaginable. Data scientists are our guides on this journey as they are creating radical new ways of thinking about data and the world around us.”

It isn’t just about pure science for science sake, either. It’s about radically increasing the usability, accessibility and business story-telling of the data now flowing through enterprises. There are proven, solid gains to be made when data science is applied to business problems. For example, the book says, data science has been shown to deliver up to 49 percent productivity gains when organizations increase data usability by 10 percent. In addition there is a potential return on assets of up to 42 percent when organizations increase data access by 10 percent. Organizations using big data to its maximum potential report a 241-percent increase in ROI, and a 1000-percent increase in ROI when deploying analytics across most of the organization.

And things will just keep getting better, as data science opens up new vistas for decision makers not contemplated before. “Advances in cognitive machine learning are on the horizon, including open source and configurable algorithms that exploit streaming real-time data’s full content, context, and semantic meaning,” Kirk Borne, principal data scientist for Booz Allen, writes in the book. “The ability to use the 360-degree view of a situation will enable the delivery of the right action, at the right time, at the right place, in the right context – this is the essence of cognitive analytics. Another way to view cognitive analytics is that, given all of the data and the context for a given object or population, the algorithm identifies the right question that you should be asking of your data (which might not be the question that you traditionally asked).

Borne adds that emerging technologies will enable organizations of all sizes and budgets to take advantage of data science, as the platforms that support it are getting cheaper and more accessible. “We anticipate advanced data science algorithms that take advantage of technological advancements in quantum machine-learning, in-memory data operations, and machine learning on specialized devices (e.g., the GPU, Raspberry Pi, or the next-generation mobile handheld ‘supercomputer’),” he writes. “In such commodity devices, we expect to see development of more embedded machine learning (specifically, deep learning) algorithms that perform time-critical data-to-insights transformations at the point of data collection. Such use cases will be in great abundance within the emerging Internet of Things (IoT), including the industrial IoT and the internet of everything.”

Here are some of the essentials of data science, as outlined by the Booz Allen team:

Data science is the art of turning data into actions. It’s all about the tradecraft — the process, tools and technologies for humans and computers to work together to transform data into insights.”

Data science tradecraft creates data products. “Data products provide actionable information without exposing decision makers to the underlying data or analytics (e.g., buy/sell strategies for financial instruments, a set of actions to improve product yield, or steps to improve product marketing).”

Data science is necessary for companies to stay with the pack and compete in the future. “Organizations are constantly making decisions based on gut instinct, loudest voice and best argument – sometimes they are even informed by real information. The winners and the losers in the emerging data economy are going to be determined by their data science teams.”

Data science capabilities can be built over time. Organizations mature through a series of stages – collect, describe, discover, predict, advise – as they move from data deluge to full data science maturity. At each stage, they can tackle increasingly complex analytic goals with a wider breadth of analytic capabilities. However, organizations need not reach maximum data science maturity to achieve success. Significant gains can be found in every stage.”

Data science is a different kind of team sport. Data science teams need a broad view of the organization. Leaders must be key advocates who meet with stakeholders to ferret out the hardest challenges, locate the data, connect disparate parts of the business, and gain widespread buy-in.”