techwatch
DOES YOUR LOGISTICS DEPARTMENT HAVE A “DATA
scientist” on staff? Probably not. But if you’re planning to
conduct a big data analysis to obtain insights into your distribution operations, you’d better be prepared to hire one … at
least on a temporary basis. “Data science skills are necessary
because the supply chain team is sitting on a lot of unused but
valuable data,” says Michael Watson, an adjunct professor at
Northwestern University and co-author of the book Supply
Chain Network Design: Applying Optimization and Analytics to
the Global Supply Chain.
A data scientist is someone trained in the
methods and techniques of extracting mean-
ing from piles of information. Generally, he
or she has a background in mathematics, sta-
tistics, and computer science. Although com-
puters and software are powerful tools for
facilitating analysis, a human expert is still
needed to make decisions about what data to
examine and how. “Data science is not a one-
size-fits-all approach,” says Larry Snyder,
an associate professor at Lehigh University
and co-author of the book Fundamentals
of Supply Chain Theory. “So you can’t just
throw terabytes of data into an off-the-shelf
system and ask it, ‘What should I do?’ It takes
data and decision-making experts to convert
raw data into useable information and ultimately, to make
decisions.”
Because so much raw information abounds in logistics,
the discipline is considered to be particularly well suited to
big data analysis. Logistics, by its nature, involves numerous
data exchanges between multiple partners to make the supply
chain flow, and there are piles of raw data sitting in all of those
partners’ systems. But it’s not just traditional data systems that
provide fodder for analysis. Big data analysis can encompass
information gathered by sensors—say, on trucks or on packag-
es in the warehouse.
The premise behind big data analysis is that if correlations
can be made between all that raw data, users can gain a bet-
ter understanding of why things happen and parlay those
insights into process improvements. “Getting to root causes
often requires analyzing data to understand
correlations—what is related to what,” says
John Hagerty, a program director for big data
at IBM.
Unfortunately, data analysis requires a par-
ticular set of skills that most logistics and
supply chain managers do not have. “The
supply chain team needs to have skills to drill
into this data and then the ability to determine
what action the company should take based
on analysis of that data—the
last part is where it is import-
ant to have a data scientist
on staff,” says Watson. “The
person would be able to sort
through the data and help
the company determine what
actions it should take or how
it should build the data into
its processes.”
Given the boom in corpo-
rate interest in big data anal-
ysis, data scientists are in high
demand right now. In fact,
according to the job website
Glassdoor.com, the median
salary for a data scientist in the United States is
currently $115,000.
That’s why companies are turning to outside
firms to hire data scientists on a project basis.
At a recent conference I attended, Gartner
analyst Michael Burkett told how one company
had to contract with an outside agency to gain
access to qualified data scientists to conduct a
big data analysis of its supply chain.
Logistics managers can expect to find them-
selves in the same situation—that is, in need
of outside expertise for their big data projects.
That’s why the first step for any manager plan-
ning such a project may be lining up an outside
data scientist for the job.
Wanted: Data scientists to
work on logistics projects
BY JAMES COOKE, EDITOR AT LARGE
Organized fo
