When it comes to understanding data and machine learning most organisations understand data in a table or spreadsheet format that is very linear, but more sophisticated relation databases are needed to make more informed business decisions and realise better WHS outcomes, according to an expert in the area.
While using spreadsheets is fine and are a good tool in themselves, they are “very lightweight” in terms of technology capabilities that are available to organisations today, said Dr. James Murray, managing director and founder of Work Healthy Australia, which specialises in developing early intervention programs and risk management strategies to help reduce lost-time injuries and proactively prevent workplace injuries.
Organisations might be able to see trend data, hotspot data, and can understand performance-based metrics, he said.
“They might have a KPI dashboard, but beyond that, it tends to get pretty thin on the ground,” said Murray.
“To use more advanced and powerful tools and be on a journey to using AI and machine learning and use this to inform decision making, organisations will need sophisticated relation databases,” he said Murray, who was speaking ahead of the AIHS Workplace Health & Safety 2020 Virtual Series: Future Safe event which will be held on Tuesday 6 October.
There are a number of common issues and challenges for organisations in this regard from a WHS perspective.
One of the more prominent challenges is siloed data sets, said Dr Murray: “they might be managing their first-aid data, WorkCover claims, and OHS procedures through separate spreadsheets or management tools,” he said.
“You need these data sets together to see the full picture. All these things are also done by humans – they are very manual.
“When you’re not digital and you are still siloing data, you don’t seize any of the opportunities.”
Dr Murray observed that many businesses are still using paper-based systems, whether it be checklists, accident and incident investigation forms.
If the forms are paper-based, and they are being turned into PDFs, he said there is no data capture, to begin with.
There are also a number of practical ways to utilise data when it comes to WHS systems and reducing OHS risks.
“First of all, you have to start with a very curious mindset. Then you have to have a question that you want to ask the data, or a problem that you want to solve,” said Dr Murray.
“The data will need to be captured and put into a relational database – there are all sorts of tools and platforms out there to help you do this.”
The next step is for the data to be “cleaned, maintained and gardened”.
“You should investigate what you need to do to improve the quality of your data, and then you can start building dashboards,” said Dr Murray.
“You then experiment with the data and start to see what can be logically predicted from it.”
It is important to understand where risk actually lies, and he said data will help with this.
For OHS, Dr Murray said risk lies with people, equipment, processes and culture.
“You can have risky people, risky equipment, risky processes and you can have a poor culture that does not recognise or deal with risk proactively enough,” he said.
“Different risks need to be dealt with individually, and businesses need to dig deeper in and analyse the risks, put strategies and programs into place and run some experiments.”
A good starting point for organisations and their OHS professionals is to understand at a base level what the technology is.
“Research it. Jump on Youtube and watch videos. Get to know how it has come to be and how it is evolving. Find out the tools and platforms that exist,” said Dr Murray.
“Then, they can map where the data is on their side of the business and come up with a plan to get that data to be visible in a way that it helps them to inform strategic decisions.”
Dr Murray will be speaking on “Data and Machine Learning: Reducing Risk with Health & Safety Systems” as part of the Workplace Health & Safety 2020 Virtual Series: Future Safe event series which will be held on Tuesday 6 October 2020 from 1:00pm to 4:00pm (AEST). For more information or to register please visit the event website.