BDCuniversity.com | BUILDING DESIGN+CONSTRUCTION | 35THEY NEED TO MAKE BETTER DECISIONS?
ity to gather current data from
the field.
Relying on physical records,
like paper forms, for data collec-
tion is steadily giving way among
contractors to custom-designed
or commercial software, a trend
that is likely to be more prevalent
over the next three years. Nearly
three-quarters of respondents
say they were satisfied with using
software as a way to collect data.
Most pros (65% of the ones
polled) still store their digital
information in on-premise serv-
ers. But 37% have turned to
third-party cloud providers, such
as Amazon, to host their field in-
formation. The benefits of using
the cloud are manifold, the most
prominent being the ability to
access data from the field while
working in the office or remote
locations, and vice versa.
Well over half of the contrac-
tors surveyed deploy mobilephone apps and cameras to
collect and send data from the
field. Expect that wave of data
to rise, as drones, sensors, and
wearables gain more traction as
jobsite tools. However, contrac-
tors are also acutely aware of
how their use of devices presents
security risks. That explains why
86% of contractors say they are
using anti-malware software,
78% are using enterprise-grade
firewalls, and 56% have imposed
policies for managing the use of
mobile devices.A PLAN TO GATHER DATA
More than one-fifth of con-
tractors surveyed have had
experience with predictive
analysis of data and business
intelligence. That compares
with just 7% for artificial intel-
ligence and 6% for machine
learning. The largest group
of respondents noted theyare aware of and understand
the concepts of each of these
emerging technologies but
have not implemented them in
their own organizations—39%
for predictive analysis, 47% for
artificial intelligence, and 33%
for machine learning.
But it is clear that technology
is advancing faster than the
AEC industry’s ability to as-
similate it. The report’s authors
offer a multi-step process to
structure a company’s data col-
lection, analysis, and reporting.- Determine what aspects of
project delivery would benefit
most from better information
to guide your data strategy. - Identify which types of
field data will best-enable
the proper analysis to gener-
ate critical decision-support
information. And think about
the minimum level of complete-
ness, accuracy, and timeliness
required for each type.- Develop a focused technol-
ogy and plan for collection and
analysis that incorporates the
specific data needed and types
of analysis required, as well as
financial and human capital in-
vestments, the time frames for
implementation, clear roles of
responsibility, and measurable
goals for success.
The report includes two case
studies—how Leander Con-
struction is using the cloud to
enhance productivity; and how
W. Soule & Co. is using data
to build better—as well as a
Q&A with Jit Kee Chin, Suffolk
Construction’s Chief Data Of-
ficer. Chin laments that what’s
keeping AI from becoming a
more efficient construction
tool is the lack of useful data.
Her viewpoint may explain why
she sees “great potential” in
automated monitoring.
as other open sources, which Krause concedes
“aren’t always the best” and might require Clark
to do more test pits or excavation “to raise our
confi dence level.”
On one of its projects, Lendlease is using
3D cameras on the jobsite that have machine-
learning capabilities that Benedict says have led
to some tangible workplace improvements (which
she wouldn’t elaborate on). Lendlease is also
compiling data to develop a tool that identifi es ar-
eas on jobsites that might be higher risk for fi res.
People movement. FXCollaborative is redesign-
ing a large lobby inside a building in New York City
to reconfi gure its traffi c fl ow. That effort includes
examining data from card readers used by employ-
ees and visitors, which told FXC what entrances
and exits are frequented most. Alexandra Pollock,
Principal, LEED AP BD+C, Director of Design Tech-nology, says this data allowed her fi rm to conduct
a pedestrian simulation analysis that will inform
where it places the reception/information desk
vis-a-vis the elevator banks to minimize the cross-
fl ow of daily traffi c.
Arup has a program called Mass Motion, which
enables users to import a virtual building design,
like a Revit model, and inject avatars to model hu-
man movement behavior. But the avatar behaviors
are based on pedestrian movement guidelines
published in the 1960s, and don’t adequately
mimic modern conditions or individual move-
ments. So Arup is developing an artifi cial intel-
ligence framework that utilizes computer vision
to gain insight into the way people move through
specifi c spaces. Algorithms are outfi tted on a
Rasberry Pi processor located next to a camera,
so people “counts” can be processed in real