A Collective Safety Net for Seafarers

Effective data analysis looks forward, not backwards

I spent more than a quarter of a century working in the maritime industry. Since coming to work ashore, I’ve dedicated my career to improving safety at sea. Reducing risks at sea starts with how we use our data, and it’s something we all need to be thinking about. Being open with our data enables us to analyse our insights more effectively. Together, data sharing and analysis empower all of us connected to the industry to make our seas safer.

Attitudes to data sharing are slowly catching up with other industries. More and more shipping companies are anonymously sharing their data with HiLo to generate life-saving insights for one another. The problem is, we still have a long way to go. In fact, when it comes to how we’re using our data, the journey has really just begun.

We’re seeing the level of shipping accidents across the industry plateau, and I think anyone connected to the industry will agree that this still isn’t good enough. It’s true that we have made progress in recent years, but every accident recorded could represent a life, an injury, environmental disaster or financial loss. We need to do more.

To make real progress we need to continue to see change. That means changing the way we think about our data and continuing to develop our attitudes to analysis. We need to do something different.

The power of weak signals

As an industry, we try to be proactive when it comes to reducing risks. The challenge we face is that there is currently no common approach to how we go about it. We often tend to look at an incident’s root cause data, instead of analysing the leading events. When we look at past events, we need to focus on weak signals to predict what may happen in the future. This approach will enable us to address those smaller failures with the aim of completely avoiding the big incident.

It’s natural to react to incidents by trying to understand their cause, hoping this will safeguard us from it happening again. To remain one step ahead of risks at sea, we need to rethink the way we use our data.

Every ship’s data is part of the puzzle

Instead of looking backwards, we need to be using our data to look forwards. The analysis of shared data enables us to calculate risks and take the right action, avoiding the incidents that are most likely to happen to our own vessels.

When a major incident happens in the industry, it can be difficult to react to it as if it could happen to one of our own vessels. We all feel concern, and we all sympathise, but because significant incidents are rare, to any one fleet at least, it’s easy to reassure ourselves that they are unlikely to happen to us. In truth, if every shipping company shared its full dataset, we would have access to a much clearer picture of the real risks facing all of us every single day.

The danger of holes in our data

One of the challenges individual companies face is that there just isn’t enough data available for them to accurately calculate risk. Because high impact incidents are rare, companies don’t generally have the information they need to understand them, and what they do have can be challenging to analyse effectively. Every voyage generates reams of invaluable data, but it’s often spread across different departments on various systems. This can make it almost impossible for staff to collate what data they do have and achieve any helpful insights from it. Holes in company data can easily lead to missing high risks, and we all know the consequences of this can be catastrophic.

Generally, historic analysis is built on the data that companies are obliged to share. This means that only the large-scale incidents and accidents tend to be used. Therefore, we often overlook small details which could help us to prevent incidents in the future. In this way, we frequently misunderstand our data. Lives are lost and ships are damaged because the data doesn’t show which events are going to be the highest risk. Unfortunately, companies waste time and money on high frequency, low risk events to minimal effect.

Predictive analytics and big data

Predictive analytics use big data collected from thousands of vessels over multiple shipping companies. To ensure that nothing is missed, information is collected from every data set, source and owner within each company. All data is collected directly from the source without sanitising, so it reflects everything that’s happened, exactly as it took place. This data undergoes statistical analysis, which identifies the likelihood of weak signals leading to high impact events and the connections between high and low impact events.

Predictive analytics demonstrate just how the small things really matter. Issues such as people entering confined spaces, such as paint lockers without proper ventilation may seem like an acceptable risk at times, but have a high risk of leading to something catastrophic – even though the data doesn’t make it obvious day to day. Predictive analytics show where seemingly safe actions can be dangerous. Using the right data and analysing it effectively empowers all of us to reduce real risks. Turning real data into tailored predictions enables you to pinpoint the most important area to focus on, rather than just those which happen most often.

As featured in Elnavi Magazine