Reducing risk of injuries utilizing safety analytics

Without a doubt, keeping workers safe on the job site should be the most important thing to any project team. And yet, today’s job sites that are increasingly large, complex, and rapidly changing make it very challenging to constantly monitor the site for all aspects of safety best practices. Having a second set of eyes that look out for safety hazards can be a valuable tool for project teams -- a key reason why many leading contractors are investing in safety training and assigning on-site safety professionals for their largest jobs. Even so, there are limits to how many site inspections teams can execute in the context of broader responsibilities that include planning and coordination, training, reporting, and incident investigations.

Mountains of data

When we take a look at all the opportunities to capture data on a job site, there are multiple collection methods and mountains of data being collected. It is not uncommon for projects to capture 1,000s of photos per month on a single project. And project teams are collecting data for many different reasons, only one of them being safety. Team members are often documenting progress, site conditions, issues, and milestones and all of these data points can be valuable to the project. Each of them is utilizing their own device and workflow to capture that information in the field. Various platforms are being fed new information constantly and there is an opportunity to learn from all of it if we can consolidate and analyze it in near real-time.

“Shane

Suffolk Assistant Superintendent documenting safety hazards.

What to do with all of this data

As Regional Safety Director for Suffolk, Marty Leik is interested in new ways to provide his teams with better tools to efficiently understand and communicate potential risks on job sites. Marty and his team at Suffolk have partnered with Smartvid.io to mine Suffolk’s huge volume of photo documentation in ways that would be impossible without technology. According to Leik, “VINNIE, Smartvid.io’s AI engine, uses a direct integration to Procore to analyze every photo we’ve taken in Procore and automatically detects every person and counts the number of times it sees things like missing PPE.” Smartvid.io finds workers missing PPE like gloves or high-visibility safety vests and brings these observations to the attention of Marty and his team. And, Smartvid.io can also identify potentially risky work environments and practices such as workers on ladders or scaffolding as well as housekeeping issues.

Why non-tech comanies are beginning to use AI at scale >

Building culture by focusing on the positive

Suffolk has implemented Smartvid.io across several jobs already which allows them to quantify and benchmark their PPE compliance from job to job in a dashboard report. While VINNIE gives Smartvid.io the information to slap workers on the wrist for non-compliance or violations, this is not the objective. As Marty says "No one responds to the negative feedback…we can now highlight where projects are doing really well." This approach is a significant driver in improving safety culture on the job site. Suffolk’s approach to driving compliance has been one of improving culture through teaching and competition. As they track PPE compliance from week to week, they measure projects against each other inciting the desire for those who are underperforming to catch up. With over 16 active projects feeding into Smartvid.io via Procore, BIM360, and OxBlue, Suffolk has plenty of data to analyze and track. Notifying workers and subcontractors that these items are being monitored and explaining to them how helps them understand you are holding them accountable for their safety.

VINNIE can watch for more than PPE

PPE is not the only thing related to safety that is identified in photos and videos utilizing Smartvid.io. The safety teams at Suffolk are also looking for trends in the use of ladders and scaffolds as well as good housekeeping that can be identified by VINNIE. They use the ability to quickly review photos of scaffolds that might have changed since their last visit and identify potential risks. Receiving notification of the presence of scaffolds or ladders on a job site doesn't always mean something is wrong, but it allows safety professionals the peace of mind of being notified and reviewing its installation and usage. As many companies have begun enforcing stricter rules based on ladders, such as "ladders last" programs, identifying photos with ladders makes it possible for safety professionals to remotely inspect job site practices and provides teaching opportunities for project teams. Safety team members and field staff can evaluate if a scaffold or scissor lift was a better choice. "It is a great way for us to push the industry to get away from the status quo," says Marty. Utilizing the actual photos and videos can provide a lot more than just a simple data point for a metric. Diving deeper into the images can identify and eliminate additional risks providing for a safer work environment. It is easy to walk by some of these hazards in the field, and the ability to further analyze what’s going on offers an opportunity for improvement. This type of workflow enables teams to not only identify lagging indicators but identify some near-miss hazards as they are reviewing the tagged images.

Overall, tracking trends in the data identified via VINNIE can help drive a more proactive safety culture on the job site. This tracking is not limited to just PPE either as we are continually expanding higher risk activities and objects which VINNIE can identify. Let Smartvid.io assist your teams in improving the identification of unsafe work practices on the job site. For more information on our system and how to get it setup on your projects reach out to one of our team members today.

Topics: Machine Learning, Artificial Intelligence, Safety



About the Author

Mike Perozek

Mike Perozek heads up Smartvid.io's sales, marketing, partner, and customer success programs. Mike is passionate about helping customers adopt new technology and achieve breakthrough financial and operational improvements. Prior to Smartvid.io, Mike built and led sales and customer success teams at leading vertical SaaS companies Buildium LLC (real estate) and Bullhorn (recruiting/CRM). Mike's alma maters include Stanford University and MIT Sloan and when he's not working he's hanging out with his family and spending time outdoors.

View more posts by Mike Perozek.

Follow me:


Leave a Comment:

Subscribe to our blog:

Using AI to predict and prevent construction incidents