There are two watchwords that crop up whenever “Big Data” gets name-dropped:
The first—and the one people tend to focus on—is transformation. The other is potential.
While the two are often used right next to one another, it is important to remember that any heralds of disruption to the very fabric of advertising, business, healthcare, et al., is predicated on a generous reading of the potential of analytics and data management technology. In reality, such potential needs a lot more help to be realized than it currently can depend on.
Computers are getting faster, smarter, and more dexterous; our collective ability to record, store, transmit, and manipulate data is growing; the applications for the mass collection of data (passively, voluntarily, secretly) are multiplying as we find more and more opportunities to replace assumptions (also known as conventional wisdom, precedent, or previous standards) with more intelligent, data-driven solutions.
In short: we have lots of data, and lots of plans to use it. The only problem is, we aren’t keeping up with the data in terms of human resources.
As Deloitte Consulting Principle Mark Schilling explained, “New technologies require people with new skills…The talent model that the CIO needs to look at and evolve becomes quite complex.”
Skeptics will look at the companies and computers and identify technical hurdles that could impede the sweeping revolution of data analytics; industry insiders are recognizing that in fact, the talent gap among people using these systems is more daunting.
Big Data doesn’t start with the analysis, and it doesn’t start with the data sets. It doesn’t even start when mass quantities of data are gathered, even if that is where it gets its name. It starts with a question, a problem, an imperfection that can benefit from the process as a whole.
As intelligent as our computers are, they are far from autonomous; they desperately need creative, capable programmers to set the parameters and determine what matters, what algorithms should be looking for, what data needs to be used. Or, as the instructors at the University of Wisconsin Data Science Program explain, “You must also be able to communicate complex ideas to your nontechnical stakeholders in a way they can easily understand.”
Data scientists need to be storytellers, capable of taking all the key characters and events and turning them into a meaningful narrative. That means knowing the audience for a story, as well as how the story ends.
Consider the hiccups afflicting the integration of analytics into medicine. In many cases, having more data was supposed to solve errors in care and organization that sprang from analog records systems and poor communication. As hospitals across the country are discovering, interdepartmental communication doesn’t heal overnight with the integration of more sophisticated records systems. Throwing iPads at doctors doesn’t help them take advantage of virtually infinite applications; without a go-between smoothing human-computer relations, communication is still the weakest link.
Data scientists certainly need to know their way around computers and numbers; but real value comes from data experts who can serve as interpreters, bridging the gap between the data, the results of crunching, and the people who need to know what it means. Big Data is just a lump of raw information until scientists can send it through processing; even so, the final product is only is valuable as the story accompanying is compelling.
The people who make the nuts and bolts of big data work are not the decision-makers; you will not see many CEOs and other leading executives tangling with algorithms. Data scientists have the unenviable task of inheriting real world needs, and turning them into questions to be asked of their data. Then, they must reverse course, take the answers they get, and make them intelligible to the people in a position to actually act upon this new information.
Soft Skills, Hard Facts
From wearable technology gathering health metrics to online shopping sites tracking our tastes, spending habits, and fashion fixations, we have plenty of spigots dumping data into an ever-growing pool.
We still need people coming up with better ways to collect data, so that we don’t forget that sheer size (the “big” of big data) can’t entirely compensate for abysmal quality, lest we make a habit of drawing the wrong conclusions from supposedly unassailable heaps of data.
And we need people who can duplicate outcomes to validate the most exciting things to come out of the big data push.
But more than anything, big data needs more than hordes of engineers and computer science students getting churned out with promises of high pay and even higher demand for their knowledge; people skills still make the difference between solving problems or simply throwing resources at them.
That, ultimately, is what makes data workers special: this is not the ascendency of the nerds, where those who work better with computers and equations than with other people are finally rewarded for their quirkiness. This is a sudden, dramatic demand for people who can work effectively with people, words and numbers.
Big data is carried on the shoulders of big promises; realizing this vast potential requires some reappraisal of who is qualified to act as interpreter between man and machine.
What is more, being well-versed in numbers and training machines to learn is core to analytics, but these skills don’t count for much if you don’t know which questions to be asking in the first place. Recognizing the systemic problems in a business, organization, or professional network that stand to benefit from the right data under the right type of scrutiny ensures that big data is actually being put to work solving real problems.
The world may be pivoting toward tech, but only so far; the rest of the pivot must be tech turning toward humans.