Technology
Photo: AndreyPopov / iStock / Getty Images Plus via Getty Images.
Rethinking
By Aatish Suman
he sudden shift of environmental, social, and governance (ESG) from an optional public relations reporting initiative into an increasingly mandatory, investor-driven reporting requirement has many companies reevaluating how they use data. While organizations have used “big data” in the past to be more competitive, efficient, and profitable, they’re now seeing the benefits of advanced data analytics in assessing EHS performance. Sophisticated use of data and analytics can reduce incidents and operational overhead, directly impacting the bottom line, improving employee morale, and strengthening the business’s reputation.
Traditional data metrics such as incident rate and lost-time incident rate only track issues after the fact. These lagging indicators reactively measure safety efforts by evaluating past performance rather than current or future conditions. While this data may be helpful for regulatory reporting, it doesn’t provide a complete picture of the cause of the issues that lead to workplace incidents or allow stepping in and disrupting the incident from occurring in the first place.
Increasingly, EHS professionals are searching for health, safety, and environmental indicators to track in order to predict and prevent incidents in the first place. The hard part is identifying the most relevant indicators to measure. That is where the combination of big data and predictive analysis becomes instrumental. Data without the proper analysis is just a boat anchor weighing everything down. Research and analysis without a statistically significant data set are not necessarily actionable or helpful.
Predictive analysis powered by artificial intelligence (AI) and machine learning — a sub-set of AI which can “learn” patterns and behavior in data — can help proactively identify areas of highest risk exposure for your employees. It can help predict and prevent workplace injuries more accurately before they occur and even uncover previously unknown factors.
The thoughtful analysis of your data can show patterns in behavior that had otherwise been obvious. By subjecting large numbers of observations and incidents to data analysis, it becomes possible to predict the likelihood of incidents of all types with surprising accuracy by pinpointing the signals that precede an event, as well as the situations in which accidents are most likely to occur.
T
Improved EHS performance
New, innovative technologies, such as sensors, remote motion capture, and artificial intelligence (AI) are making it easier than ever to collect, manage and analyze leading EHS data to predict and address areas of risk and inefficiency. Predictive analysis can focus your activities most likely to happen in the future.
Companies that implement big data and machine learning solutions can benchmark themselves, compare performance against others in their industry, see data correlations, and receive predictive and prescriptive insights that help improve EHS operations. Machine learning-based solutions eliminate much of the guesswork around the data collected to monitor operations and find patterns to help predict and prevent failures, leading to significant cost savings, higher predictability, and increased availability and use of the monitored systems.
My studies have proven the increasingly common view that machine learning in predictive maintenance outperforms traditional maintenance strategies. It’s both an attainable and useable tool all companies should implement to achieve operational excellence. Whether used to analyze incident reports for categorization, facilitate efficient root cause and corrective action identification, extract requirements from regulatory documents for auditing purposes, or use computer vision for remote ergonomic evaluation and analysis, machine learning replaces the costly human labor efforts required to complete these tasks and significantly increases safety, time/cost efficiencies, productivity, and profitability.
Implementing the change
Technology is rapidly changing how companies approach EHS. As organizations become more aware of the benefits of predictive analysis and machine learning, EHS software is developing better ways to handle big data quickly and effectively. It’s essential to look beyond the hype of such systems and focus on the fundamentals of their operation. The software should track and analyze data but also be simple and easy to use. Software should also be conclusive; a system that’s limited to a small number of factors is inherently less powerful than one examining a long list of workplace indicators.
In addition to data entered by safety professionals, the software should also track behavioral factors and information about the environment within which workers perform tasks. That is why, ideally, all employees should be contributing data for analysis. Mobile applications and web apps provide an easy way for workers to input data automatically uploaded and ready for analysis. The ease of mobile also helps ensure the data entered is more accurate, as workers upload the details right away instead of waiting until after their shift has ended or relaying the information to someone else to submit.
An effective predictive analysis system should do more than provide a high-level view of the potential for accidents. It should allow the users to drill down to specific tasks and individual workers to identify the areas with the highest risk. While it’s helpful to say that workers in one facility area appear to be at higher risk, it’s far more useful to identify a worker more likely to be injured when performing a particular task under specific conditions.
As organizations become more aware of the benefits of predictive analysis and machine learning, EHS software is developing better ways to handle big data quickly and effectively.
Proven benefits
When companies consider investing in the internet of things (IoT) as part of the Industry 4.0 revolution, it is crucial to investigate the actual cost of your solutions. Powerful visioning tools allow data collection necessary to drive AI and machine learning from video sources, eliminating the need for hard-wired devices. For instance, workers are no longer required to put on wearable sensors to perform motion capture assessments. Sophisticated software like VelocityEHS Industrial Ergonomics enables anyone with a mobile device camera to record an employee doing an activity, run it through the software, and generate an expert performed ergonomics assessment within minutes.
Sensorless motion capture provides considerably higher accuracy than visual observations. While it is somewhat more accurate at identifying postures than experienced ergonomists, it’s substantially more accurate than less technical assessors. From a practical standpoint, this means that an inexperienced ergonomics team member using sensorless motion capture can assess a job with similar — and often better — accuracy than an ergonomics specialist, which is its most important benefit. Among ergonomics teams to design engineers, this method supports a participatory approach to ergonomics, involving and engaging many and resulting in a visible and sustainable process.
No matter how experienced or dedicated an assessor may be, it’s hard to collect accurate information about the percentage of time an operator spends in awkward postures. Although this is an essential factor, most observational tools don’t include more than two levels of exposure for this variable. However, sensorless motion capture provides the ability to accurately capture this data and apply multiple levels of exposure, which results in better prioritization of jobs. It helps obtain greater fidelity and gather more data than the human eye can see while also eliminating the variability and human error that comes with traditional, observation-based methods.
While the time savings per assessment may be relatively small, the cost savings per assessment is substantially higher, as a much wider range of people in the organization can accurately complete an assessment. The cost savings for an assessment can exceed 50 percent. What can’t be quantified is the value of replacing competing opinions with factual evidence. How much time has your ergonomics team wasted debating the risk scoring of a job? While the total amount of time savings is modest, the ability to have increased accuracy and detail combined, with even a small decrease in time required, is beneficial.
The future of safety Is now
Committing to an EHS data analytics and performance management process can provide significant financial and operational returns. When companies and their EHS leaders can analyze data to predict issues in advance, make corrections as quickly as possible, and better ensure workplace safety, they’re applying the world-changing promise of ESG.
While developing and implementing a predictive analysis and machine learning strategy can seem intimidating, it has the potential to enhance business efficiency, improve safety outcomes and lower costs related to your workplace safety initiatives. The shift is here to stay, don’t miss the opportunity to drive your business forward.
Aatish Suman is a machine learning scientist at VelocityEHS, the global leader in cloud environmental, health, safety (EHS), and environmental, social, and corporate governance (ESG) software. To learn more about the company’s innovative machine learning technologies, visit www.EHS.com.