What's Your Problem?

• DOI: http://doi.org/10.1002/jcaf.22085
• Date: 01 September 2015
• Published date: 01 September 2015

95

© 2015 Wiley Periodicals, Inc.

Published online in Wiley Online Library (wileyonlinelibrary.com).

DOI 10.1002/jcaf.22085

C

o

m

m

e

n

t

a

r

y

What’s Your Problem?

Tim Chartier

Data analytics is an emerg-

ing, if not already pres-

ent, field of business.

Iwork in data science collabo-

rating with businesses, profes-

sional sports teams, and college

organizations at Davidson

College, where I teach. What

lessons can we learn about

problem solving from this rela-

tively new field? In this article,

we look at the importance of

having a clear vision of the

problem you wish to solve or

question to answer.

At first glance, data science

may seem like a field where one

need not define a precise ques-

tion. While this may be true at

times, it is generally far from

the rule. For example, at the

beginning of the 2013–2014

season, the NBA installed

SportVU cameras in all of

its arenas. The technology

captures x‐y coordinates of

every player on the court and

all three spatial positions of

the ball. Such information is

captured 24 times per second

for the entire game. This leads

to a massive amount of data

containing tremendous infor-

mation on the game. With such

raw data, as it is often called,

new insights emerge from Kirk

Goldberry’s heat maps showing

the frequency of shots to new

work introduced at the 2015

MIT Sloan Sports analytics

meeting on identifying defen-

sive match‐ups from such data.

From such examples it can

appear that large data means

big insights. While NBA teams

have analysts, many note the

continued difficulty of min-

ing such raw data. Goldsberry,

for example, approached the

creation of heat maps from the

view of his field—cartography.

His question grew from his

training: How could the

data give spatial insight on

shooting?

If one does not have a

question in mind, then tackling

a large problem isn’t like look-

ing for a needle in a haystack.

While difficult, one can devise

methods to look for a needle

given the knowledge of the

goal. Without a definite goal,

one sifts through the haystack

hoping to discover something.

Yet searching for an easily seen

large gem versus a stealthy nee-

dle is quite different. As such,

defining the goal can define an

appropriate process.

Yet choosing a process can

also be guided by careful and

continued focus on the goal.

I illustrate with an example

from my research group. We

were recently posed an open

problem from a professional

sports organization. I teach at a

highly selective liberal arts col-

lege where many students desire

independent research projects.

For this problem, the data

set was large and could answer

many questions. Our job was to

select relevant portions of the

data and analyze them to come

to conclusions about the posed

question.

After a week, my students

were moving forward but at a

much slower rate than I antici-

pated. In the first of our two

weekly meetings, I learned that

the students were integrating

the data into a database. I

thought, “Oh, of course. Why I

didn’t think of that? It is the

natural step that everyone takes.”

The students noted that such

work would give flexibility to

future problems.

The next week we had

made progress, but again at a

much slower rate than I

originally anticipated. Such

a miscalculation regarding

the time needed for research

isn’t uncommon. As Einstein

said, “If we knew what we

were doing, we wouldn’t call it

research.” As such, we may not

know the exact process of solu-

tion until the problem is solved.

Yet the difference between my