Data and analytics in the EHS world can be intimidating. But, when even OSHA starts talking about the value in leading indicators, we have to assume this is coming. Ready or not.
Last week I had dinner with a friend who is a “Big Data” guy and we got into a discussion about analytics in the EHS world. My general sense is that for many in our profession, the feelings of the four friends in the forest in the Wizard of Oz are pretty familiar: the jargon alone surrounding data analytics can be very intimidating.
However, when even OSHA starts talking about value in leading indicators, we have to assume this is coming, ready or not. It is worth talking about why the topic makes so many uncomfortable. I see three key areas worth exploring:
1. Collecting Good Data
First, to get to leading indicators, you have to have data, and probably lots of it. A problem here is that there is good data and bad data. Bad data is data that is inconsistent, contains errors or is just missing.
There is also useful data and unhelpful data. If I am trying to understand the impact of airborne particulates on my workers in my packaging department, then it isn’t helpful to measure the outside temperature – or is it? Could there be some kind of connection between cold weather outside and effectiveness of the filtering systems inside? Maybe – and that becomes both the opportunity and burden on analytics initiatives.
How do I know what data to collect? Gathering and storing data is not cheap, so it is helpful to know what specific questions you want to answer in order to know where to look and what to collect.
2. Make Sense of the Data
Second, the tools available today to make sense of data, to find correlations or connections, are a big leap from Excel spreadsheets.
We can all muddle our way through rows and columns to find totals, percentages and line graphs. Creating your own statistical analyses using Tableau or Power BI is a horse of a different color. Fortunately, there are companies out there who specialize in this kind of work if the questions for which you seek answers are important enough.
Also, some software vendors can be helpful in guiding you through basics of data analysis using their tools. Take comfort in knowing that like most technology, you can count on these tools getting simpler over time.
3. Next Steps
Third, even when we have our data collected and our dashboards ready, what do we do now? You can certainly automate preparation of reports required by your management group, saving time every month. But the real value of this powerful reporting comes from identifying opportunities for process improvement.
An example from daily usage by many of our customers comes from wanting assurance that DVIRs are being properly documented on a daily basis and over time. By documenting DVIRs digitally, a real-time chart will tell you at a glance if they are being completed, where and by whom.
Do you want to know if there are patterns of inspection failures in these DVIRs? No problem, as the same dataset has the answers that could lead to improvements in maintenance practices. It gets more complicated, but potentially far more valuable, when you find correlations that may not be at all obvious. How about our airborne particulates example above – what if it turns out that there is a direct correlation between particulate levels inside and the level of raw materials in feeder hoppers when outside temperatures are below 40° and it is raining.
The point is that each of these concerns are both legitimate and addressable. One good starting place to expand your understanding is a Google search for: “beginner’s guide to data analytics in operations”. This may seem like a lot to take in, but remember the response to the old question, “How do you eat an elephant?” Answer: One bite at a time.