Click Here for a Data Scientist: Big Data, Predictive Analytics, and Theory Development in the Era of a Maker Movement Supply Chain

Published date01 December 2013
Date01 December 2013
DOIhttp://doi.org/10.1111/jbl.12024
Click Here for a Data Scientist: Big Data, Predictive Analytics, and
Theory Development in the Era of a Maker Movement Supply
Chain
Matthew A. Waller
1
and Stanley E. Fawcett
2
1
University of Arkansas
2
Weber State University
Predictive analytics is impacting many diverse areas, ranging from baseball and epidemiology to forecasting and customer relationship man-
agement. Manufacturers, retailers, software companies, and consultants are creatively discovering new applications of big data using pre-
dictive analytics in supply chain management and logistics. In practice, predictive analytics is generally atheoretical; however, we develop a
292 model to explain the role of predictive analytics in the theory development process. This 2 92 model shows that in our discipline we
have traditionally taken one path to theory development, but that predictive analytics can be a salient component of a comprehensive theory
development process. The model points to a number of research questions that need to be addressed by our research community. These ques-
tions are not just highly relevant to the academic community but also in urgent need of answers to help practitioners execute the right strategies
with greater precision and efciency. We also discuss how one disruptive trend, the maker movement, changes the nature of who the producers
are in the supply chain, making big data even more valuable. As we engage in higher levels of dialogue we will be able to make meaningful
progress addressing these vital research topics.
Keywords: data science; predictive analytics; big data; data scientist; maker movement; 3D printing; additive manufacturing; theory
development; supply chain management; logistics; education; research
He gets on base a lot. Do I care if its a walk or a hit?
Billy Beane
INTRODUCTION
1
The epigraph, a quote by Brad Pitts character Billy Beane in the
2011 movie Moneyball, sums up the Oakland As unconven-
tional approach to building a winning baseball team. As told in
the book Moneyball: The Art of Winning an Unfair Game by
Michael Lewis (2003), the Oakland As used predictive analytics
to win more games with less money. The idea is increasingly
familiar. Long-held approaches to doing something may not be
the best way to get the job carried out. For the Oakland As, this
meant moving away from traditional metrics like RBIs, batting
average, stolen bases, and defense metrics because they are not
predictive of success. Out of desperation, the As used predictive
analytics to identify which metrics really made a difference. With
this knowledge in hand, the As build a competitive teamwith
about a third of the budget of the Yankees in 2002. The As
made the transition from losers to winners and reached the play-
offs, dramatically increasing the teams value. The rest of major
league baseball soon followed, hiring data scientists to build
more competitive teams. The key learning point: The As did not
pick the independent variables ex ante based on theory. They
picked the vital decision criteria based on statistical associations
among the variables.
Others have shown that the enabling power of leveraging big
data with predictive analytics is not a uke. Consider the follow-
ing two examples.
Google sought to predict the timing of u outbreaks geograph-
ically based on search term frequency (Mayer-Sch
onberger
and Cukier 2013). Googles model was completely empirical,
not based on theory or even logic. You may be thinking,
Google got lucky. A more theoretical approach in the hands
of experts would have worked even better.The Centers for
Disease Control and Prevention pursued the more traditional
approach, developing models based on biology and epidemio-
logical theory. But, the Google model outperformed the
CDCs models, more accurate at predicting H1N1 outbreaks
in 2009 (Mayer-Sch
onberger and Cukier 2013).
Walmart contributed to folklore among disaster relief experts
by studying point-of-sale data from stores that were in areas
where hurricanes were imminent. Logic would suggest that
facing disaster, people would buy saws, shovels, and safety
equipment. But, the data showed that people buy Pop-Tarts,
among other things, in unusually large volumes. This was not
deduced with the use of a theory but through correlation
analysis. This unusual insight turns out to be useful for fore-
casting and inventory management prior to a hurricane. That
is, data-driven insight reduces the cost of lost sales and other
logistics costs, in addition to supplying hurricane-stricken
areas with sufcient Pop-Tarts and water.
1
Our earlier editorial, Data Science, Predictive Analytics, and
Big Data: A Revolution that will Transform Supply Chain
Design and Management(Waller and Fawcett 2013), dened
data science, predictive analytics, and big data, within the con-
text of logistics and supply chain management (SCM).
Journal of Business Logistics, 2013, 34(4): 249252
© Council of Supply Chain Management Professionals

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT