Realtor - September/October 2013

It’s hard to avoid conversations about big data these days.

The concept has become politically charged amid debates on just how the federal government and large corporations are using the vast stores of data they’re gathering on us. In areas that range from thwarting terrorism to understanding consumer tastes in toothpaste, there is vast disagreement about the ramifications. Some see unprecedented threats to personal privacy; others welcome the potential for increased e;ciency in decision-making. Most of us are unsure what to think.

Strictly speaking, big data doesn’t just mean “a lot of data,” though that’s certainly one element. Nor does it refer to large organizations collecting data, though that too tends to be an attribute. And while it’s sometimes employed in both a pejorative way—in the same spirit as, say, “Big Government” or “Big Oil”— and in a positive but overly generalized way

By Brian Summerfield

akin to “cloud computing,” those characterizations are more misleading than elucidating.

Big data refers to the increasing interconnectedness of data, the melding of large, complex information sets that are nearly impossible to process using traditional database management and processing tools.

The promise of all that piling on lies in the ability it gives researchers to predict human behavior. Most famously, big data has been used to draw a connection between rising food prices in the Middle East and the Arab Awakening; theoretically, it might someday help you determine when people will move and what type of home they’re likely to buy.

Deafening Buzz

Even in its nascent stage,
in real estate and other
disciplines, the phrase big
data is being used in ways
that threaten to render it
meaningless. It’s bandied
about on Twitter and
Facebook; dissected and
discussed at industry con-
ferences; and increasingly
used to pitch marketing
and information manage-
ment solutions that can
purportedly help brokers
run their businesses bet-
ter. It is, as Realogy Presi-
dent and CEO Alex Per-
riello sardonically put it,
“the silver bullet of 2013.”
“Most people think
it’s just the same old data
processing as in days past
but more of it,” says Ken-
neth Cukier, data editor
for The Economist maga-
zine and author of the
book Big Data: A Revolu-
tion That Will Transform
How We Live, Work, and
Think. “That’s not true.
Big data isn’t just how
people interact with lots
of information, but about
how computers can pro-
cess vastly more data to
do new things.”
Perriello adds: “Big
data is really the amalga-
mation of separate but re-
lated data sets. When you
aggregate them, they pro-
vide valuable insights that
you can’t get by looking at
them separately.”
In practical terms,
that means business intel-
ligence is getting more,
well, intelligent. “By tap-
ping vastly more data, we
can ask new questions
and do new things,” Cuk-
ier explains. “And we’re
harnessing types of infor-
mation that we never had
before. With this, we can
find correlations that be-
fore escaped our notice.”
Large retailers such
as Wal-Mart and Ama-
zon are already using big
data to great e;ect. These
companies are finding
patterns among con-
sumers and using those
insights for target mar-
keting and product po-
sitioning. (If you’ve ever
received a “You might
also like …” message after
purchasing something
online, then you’ve gotten
a small taste of big data.)

“In economics, when
we improve transac-
tions, price setting, and
liquidity, all parties gain,”
Cukier says. “Big data
makes markets more e;-
cient—so everyone stands
to benefit.”
And what holds true

A Boolean-Bayesian Bouillabaisse

For data geeks, the rise of big data represents a transition from the Boolean to the Bayesian. Before you shake your head in dismay and turn the page, here’s a simple explanation from the NATIONAL ASSOCIATION OF REALTORS®’ Chief Technology O;cer Mark Lesswing.

Business intelligence and technology used to be dominated by Boolean logic, named for 19th-century English mathematician George Boole. Essentially, this refers to a form of algebra that’s binary in nature and represents a pair of di;erent, absolute values. (Think “yes or no,” “true or false,” “black or white,” “ 1 or 0,” and so forth.)

Today, business intelligence is considered in terms of Bayesian probability. Termed for another long-dead English mathematician named Thomas Bayes, this involves conditional logic that can change based on the information streams. If the visual structure of Boolean math is two parallel lines, then Bayesian math is like a star, with multiple streams of data meeting in the center to produce a conclusion. With the latter, you can combine data sets in practically unlimited combinations, and if you do it correctly, new insights about consumer behaviors can emerge.