On Jan. 29, 2014, Safety+ Health, a magazine produced by the National Safety Council, hosted Predicting & Preventing Workplace
Injuries: A How-To Guide. The webinar addressed how companies can use
safety data and analytics to predict and prevent injury.
According to speaker Griffin Schultz, the general manager of
Predictive Solutions Corp., there is a new paradigm for data analytics
consisting of not only collecting and analyzing data, but also using that data
to predict future outcomes. This paradigm is supported by technological
advancements that allow computers to predict future events according to data.
Basic data analytics generally involve collecting and
organizing data that can then be used to answer questions such as, What happened? Where did it happen? and How
often did it happen?
Advanced and predictive analytics, however, take analysis a
step further, providing answers to more complex questions, such as, Why did it happen? Will trends continue? and
What might happen next?
Companies can get started using predictive analytics by
utilizing software solutions such as SafetyNet, a safety management service
that is being used to gather more than 1.8 million workplace safety
observations per month. Customers enter their own inspection data, which is
then compared to other inspections not only within the same company, but also to
data entered by other users of the service. Those comparisons identify
correlations between inspection elements and negative outcomes, and SafetyNet’s
red flag model allows users to see which of their current projects, sites or
work groups may be at risk.
Schultz warns that a company cannot rely solely on this
predictive model to keep workers safe. He likens the red flag SafetyNet
displays next to at-risk operations to the check engine light on a car. The red
flag helps alert a safety professional to a potential hazard, but additional
inspections and basic data analytics then need to be performed to determine
what element of the operation is causing the risk and how that hazard might be
mitigated.