Big Data and the Prediction of Workplace Accidents

Big data offers employers an unprecedented opportunity to move from lagging safety management into predictive models.

What is big data?

To put it simply, big data is larger, more complex data sets. With big data, the data collected is so large and so unstructured that traditional computer processing cannot manage it. Modern analytics platforms now have the capability to take unstructured data collected by cloud connected devices and organise it into more digestible datasets, on a real-time basis. 

Which workplace technologies collect data?

Most emerging technologies have the ability to collect large volumes of raw digital data. From drones to wearable devices, and machine sensors to robots, the Industry 4.0 technologies we are adopting into the workplace are now providing us with precious and usable information.  This data and information can assist safety departments in monitoring workers, work equipment, and the work environment. Not only can companies use the data to understand why a worker fell, they can also use the data to predict what workers may fall next, and where they may fall. This is the data’s true power; its ability to present the safety manager with actionable information for decision making through the power of “predictive analytics”.

Predictive analytics and accident prevention

Predictive analytics make predictions about the future. These future events can be anticipated using statistical techniques such as data mining, analysing current and historical data, and predictive modelling. 

There is huge opportunity for safety critical industries such as mining, heavy construction, and manufacturing to monitor workers through data-collecting technologies such as wearables. For example, a mining worker’s vitals may be monitored over a period of months to understand their peaks in fatigue, which has negative effects on judgement and performance.

With this information a health & safety department will be able to gain insights into the times of the day the worker is over-fatigued and likely to make poor judgements

Armed with this information, employers can advise a worker on the work shifts and roles that suit them personally – and when they should factor work breaks into their day to lower the risk of fatigue related incidents.

The same general process can be applied to machinery and work equipment. Safety departments can, via IOT sensors and equipment telemetry,  liaise with maintenance and engineering departments to ensure that key equipment stays in optimal repair – and does not fail during operation.

Potential benefits for industry

The use of predictive analytics has huge performance benefits for occupational health and safety. Leveraging data to predict accidents is being driven and developed by some of industries biggest enterprises. However a recent survey carried out by the National Safety Council revealed that only 12% of safety critical industries in the US were utilising predictive analytics to anticipate future events. This leaves huge scope for the technology sector to develop systems and tools to further grow this space. 

Progress and challenges for data analytics

A general rule with data analytics is ‘the-more-data-the-better’ for predicting future patterns. All data may not be relevant to each business division however a unique opportunity exists where departments can share employee data. This would allow both worker performance and worker behaviour to be optimised, gaining a sharper view of the ‘factory floor’. There is also a multitude of secondary benefits such as cost efficiencies due to accuracy of resource application.

This is all heavily dependent on worker buy-in regarding the use of wearables, and the utilisation of their personal data by third parties. Workers will rightfully be skeptical of the use of their health information for work purposes, and the security of their personal data thereafter. This privacy aspect of data analytics is key to the success of leveraging data to move occupational health & safety away from lagging metrics, and into a more predictive model. Developers who can find solutions to this privacy-related conundrum will see unprecedented demand for their products and software. 

Garry McGauran is author and editor at Emerging Tech Safety. He has 17 years experience as a prototype risk assessor, design safety consultant, and academic research advisor, as well as heading up his own drone inspection service. He is a freelance safety consultant serving the tech, industrial, and utility sectors in Ireland and the UK.