Shyam's Slide Share Presentations


This article/post is from a third party website. The views expressed are that of the author. We at Capacity Building & Development may not necessarily subscribe to it completely. The relevance & applicability of the content is limited to certain geographic zones.It is not universal.


Tuesday, April 30, 2013

Talent research is redefining the way companies identify and predict the core capabilities that make a data scientist successful. 05-01

Talent research is redefining the way companies identify and predict the core capabilities that make a data scientist successful.

Fifty-five percent of big data analytics projects are abandoned.

This surprising finding comes from a recent survey of 300 IT professionals, conducted by a company called InfoChimps.

The most significant challenge with analytics projects, according to the survey? Finding talent. Most (80%) of the respondents said that the top two reasons analytics projects fail are that managers lack the right expertise in house to “connect the dots” around data to form appropriate insights, and that projects lack business context around data.

Greta Roberts, CEO of Talent Analytics Corp. says that part of the reason there is such a skills shortage with data scientists is that the current job description, often the one floated by Thomas Davenport and D.J. Patil, doesn’t quite hit the mark.

“It’s over-specified,” said Roberts. “There is a null set of people that fit the entire description. They’re unicorns; you can’t find them. Or there are a very limited number of people that fit the criteria.

“When you review data scientist hiring criteria you’ll find mutually exclusive requirements,” Roberts continues. “They want charismatic communicators that are able to effectively present findings. At the same time, they want people to sit and work with data all day. These are two different types of people. Our data shows companies in fact split up these roles.”

In the October 2012 issue of Harvard Business Review, Davenport and Patil popularized the idea that data scientists have “The Sexiest Job of the 21st Century.” These folks, they suggest, can do it all: make discoveries, write code, understand their technical limitations while fashioning new tools, conduct academic-style research and communicate effectively.

Roberts isn’t so much criticizing the work done by Davenport and Patil — both are leading researchers in the area of data analytics — as she is expanding upon their definition of a successful data scientist. As a faculty member at the International Institute for Analytics where Davenport is a co-founder and research director, Roberts’s team conducted research to determine if there is a common “fingerprint” among all data scientists. They looked for characteristics that are different from skills, experience or education — traits that govern motivation, indicate creativity and drive success.

Roberts’ research showed that there is a clear, measurable fingerprint. Published in December 2012 — Roberts presented these research findings at Predictive Analytics World in mid-April — Benchmarking Analytical Talent outlines those characteristics that make up a data scientists, beyond education and skill set:

They have a cognitive “attitude” and will search for deeper knowledge about everything.
They are driven to be creative and will want to create not only solutions, but also elegant solutions.
They have a strong desire to “do things the right way,” and will encourage others to do the same.
They have an extremely high sense of quality, standards, and detail orientation, often evaluating others by these same traits.
They tend to be somewhat restrained and reticent in showing emotions, and may be less verbal at team or organizational meetings unless asked for input or if the topic is one of high importance from their perspective.
They may take calculated, educated risks — only after a thoughtful analysis of facts, data, and potential outcomes. They persuade others on the team by careful attention to detail, and through facts, data, and logic, not emotion.

The findings, Roberts believes, can help organizations suss out qualified data scientists (or sales people, call center reps or other professionals) by identifying characteristics of raw talent, and then creating a benchmark that others can be measured against.

So what does this all mean for hiring managers? Perhaps more important, where are organizations getting it wrong?

“A popular approach is to hire for skills,” said Roberts. “You’re going to have a lot of failures if you just say ‘I need SPSS, R, SAS’ or some other skill. Business and technology are evolving so fast now. You need someone [who] is compelled to learn and keep up with what is new. So, it’s the curiosity to learn the skill that is the fingerprint. Not the skill itself.”

Creativity and curiosity, she says, are far more important than established skills.

Another misstep is not recognizing the difference between candidates being curious or just detail-oriented — both very different attributes. The way to determine the difference? Asking questions that get at curiosity.

Roberts also says organizations often focus too much on PhDs, a credential that may mean a candidate is more suited to research, rather than performing the required actions in a commercial setting.

Finally, HR managers should be aware that analytical professionals are just that — analytical. As a general rule, they are not likely to be charismatic and may not present well in an interview.
Reproduced from MIT Sloan Management Review