Database Reference
In-Depth Information
When a reputable business magazine proclaims the “Sexiest Job of the 21st Century,”
it is certain to attract attention. So it happened in 2012, in a
Harvard Business Review
article written by Thomas H. Davenport and Greylock Partners' data scientist in
residence, D.J. Patil.
3
According to the authors, this sexy job is—surprise!—the data
scientist. Despite the assumed sex appeal, some analysts predict that demand for this
role will outpace supply. A McKinsney report claims that by 2018, the United States
will be short by at least 140,000 of these sexy data scientists. Who are these important,
fetching people, and why do we need so many?
Like many technological buzzwords,
data scientist
can mean different things to
different people. Because of the current state of data technology, people who are suc-
cessful in this field often have to wear many hats. They must possess an understand-
ing of systems and software necessary to collect and extract the data they need. They
will often have at least enough statistical and measurement knowledge to understand
whether they are asking the right questions.
So what exactly are these people? Are they developers? Are they statisticians?
Davenport and Patil assert that a data scientist is a person who possesses a combina-
tion of intellectual curiosity, domain knowledge, and the technical chops to solve data
challenges. The Davenport and Patil article makes a good point that many organiza-
tions already have people like this, perhaps even stretching out of their current posi-
tions as developers or statisticians to solve problems. By formalizing and evangelizing
the data-scientist role, people like Patil hope to call attention to the success that these
contributors have in organizations.
Sean Taylor, at the time a Ph.D. student at NYU, posted an interesting article on
his blog called “Real Scientists Make their Own Data.” One of Taylor's assertions
is that a problem with the term data scientist is that it can be applied to people who
are not, in fact, scientists. In Taylor's view, true scientists are people who take the
time to build their own datasets. There is something very human about the role of a
traditional scientist. Scientists tell stories and then use experimental data to provide
convincing evidence for their view. Do data scientists do this as well? Science is often
concerned with establishing basic principles through observation and experimentation.
Data scientists aren't researching the basic principles of data. Data scientists tend to do
the kind of applied, practical work that we normally think of as engineering.
It may come to pass that the term data scientist turns out to be a poor one. A more
useful question might be, “What are the skills that make people in this role success-
ful?” Proponents of data analysis technology often debate whether or not the com-
ponents of the skill set necessary to collect, analyze, and manage massive datasets is
something that can be automated. Some claim that the skill sets necessary to deal with
current data challenges comprise a new type of job category, one that will be around
Search WWH ::
Custom Search