From data to intelligence
Using data-led insights to unlock efficiencies and improve performance.
Data collection and analysis is often celebrated as providing a unique route to improving performance and making cost savings. And the more data we have, so we are told, the better. That’s fine in theory, but it isn’t always clear what this means in practice, or how it can help with the big strategic decisions rail leaders face.
So let’s start with some big numbers about rail. The mainline railway in Great Britain has 9,864 miles of route, on which over 10 billion tonne-miles are carried by freight each year. In 2022 there were almost 1 billion passenger journeys a year and over 294 million passenger miles. All this is compressed into a country that is just under 600 miles long.
We know that, potentially, there is a wealth of different data sources, as well as huge volumes of data, within rail that could reveal the real way rail operates and real issues that need attention.
But are there data insights that could transform rail? Well, the potential treasure trove of data is extremely varied, and different datasets don’t always fit together easily. For instance, a recent analysis of depot safety combined disparate data from a large range of sources: SMIS, TRUST train movements; ORR depot stewardship scores; weather data; depot utilisation data; and workforce breakdowns of hours and shift times. Combining data from these multiple different aspects of operations can enable analysis that reflects the complexity of how the railway actually operates.
Combining data, and at scale, enables the detection of small but significant causal relationships between different factors, causal relationships that would otherwise be overlooked. Revealing these enables leaders to pinpoint the areas for effective action to improve performance and efficiency.
A few examples show the benefits of this approach:
Combining SPAD data with data about summer weather revealed that SPADs increased slightly during periods of extreme heat.
Analysis of depot safety data revealed that accidents are most likely around 10am when trains return from rush hour.
Combining the analysis of rainfall patterns and earthwork failures has provided data insights for our Whole System Risk Model which minimises the use of disruptive blanket speed restrictions following extreme rainfall.
As well as revealing what the problem is, when it occurs, or the multiple solutions that may be needed, data insights can also help identify who rail needs to work with to solve a particular problem. For instance, data about objects on the line can be categorised into various topics which includes ‘animals’, and the animals category can be further broken down into type of animal. If the animal is a farm animal, there’s a farmer who can be worked with to reduce future risks at that location. By contrast, if the animal is a deer, it’s probably wild and other solutions will need to be explored to reduce future risk. RSSB’s data insights team therefore has a rich world of insights to explore.
In addition, the team has a wealth of analytical techniques and approaches to use. Machine learning, statistical models, and data visualisations enable the team to convert data into meaningful insights and model scenarios that can be used for decision making.
But our valuable data insights aren’t only based on technical prowess; sector knowledge and collaboration helps identify insights wherever possible for the benefit of the whole industry. Whatever the precise reason, through data insights capabilities rail can continue to thrive.
View page
