Joining Chiltern in April, in the teeth of plummeting passenger numbers due to the Coronavirus pandemic, Loic Liegeois, commercial insight at Chiltern Railways, sees data as a key element in understanding the changed approach of passengers to rail. Quadrant Transport sits down with Loic to find out more
Having joined Chiltern Railways just six months ago, commercial insight manager Loic Liegeois is already a key cog as part of a digital team eager to bring a data-centric mindset to the company and the wider rail industry.
Part of the TOC’s Data Ecosystem project, seeking to push the capabilities of data visualisation for its own team members and customers, Loic is tapping into areas to spread the word of the value of data: in passenger count modelling; leveraging datasets through machine learning; and building a cultural shift in mindsets towards data from Chiltern employees.
“There are lots of different things to talk about and demonstrate in the environments that we’re trying to store our data in,” Loic told Quadrant Transport.
One of the flagship projects from the wider Data Ecosystem agenda from Chiltern is that passenger count modelling system. In understanding the sharp fall in passenger numbers using public transport following the Coronavirus pandemic, Chiltern is utilising a variety of data sources – gate-line data, CCTV passenger counts, and advanced ticket yield income – to collate and utilise for transport managers at Chiltern to understand the wider landscape of passenger numbers.
“That’s the extent of the focus that’s being put on passenger counting at the moment,” said the commercial insight manager Loic. “If we know how busy our trains are, then we can start to think that – if we now know that this particular train is actually always empty – then you can start to provide cheaper tickets here, because otherwise we’re just wasting empty seats.”
If you don’t have someone at a high enough level saying that we would benefit from automating this data-feed, then you’re not going to get anywhere
Loic continued that these real-time streams of data could allow train units to be arranged to accommodate changing capacity in response to volatile passenger numbers, or perhaps utilise the national timetable provider, Darwin, to feed live passenger numbers into timetabling systems across the network to make available for customers, for example. “The end state we want to be able to get to is to be able to monitor the entire network, and every single train,” he claimed.
The data message – visually and culturally
Of course, the Data Ecosystem project has clear goals for improving the business intelligence of Chiltern. The commercial insight manager said that his team seeks to create dashboards which produce a variety of data visualisations in the cloud – the passenger count model being the first piece of analysis – that various departments at Chiltern can access and utilise as to they see fit.
“The next step is to add in our actual train formations,” Loic explained. “Once we have our train formations into the data platform itself, we can understand which trains have been short-formed, for example, and then you can make further decisions on that.
“It’s really amazing, because we can start to build onto it. That’s the benefit of this cloud environment, where we can start to bring in datasets and enhance that, and keep working on it, rather than it being just one single siloed thing. That’s a big part of the project.”
That’s an example of what data can do on the solutions-focussed, customer side of the business, but what about staff? A report by the World Economic Forum found that 85% of companies worldwide would need data analysts by 2022 – but Loic says the rail industry needs to go much further than just employing young analysts to drive data-centric decision-making.
“There’s quite a few challenges [to encouraging data usage in rail],” says Loic. “The biggest challenge is understanding the use case and the lack of understanding of the use case itself. When people think data, they think Excel spreadsheets, they think: ‘let’s just sum this up and turn it into a graph’.
“It’s about changing that mentality – so one of the things that I started doing when I joined Chiltern is just demonstrating little things. you can look at the outcome, and instantly you show someone, for example, how quickly you can segment your customer sets. You’ve shown them how quickly you can segment your customers, and how you can do it without applying bias purely by running an algorithm – it does it in two seconds,” outlined the commercial insight manager.
Whilst highlighting use cases and showcasing examples of data analysis can help getting the message across, just speaking to people and their challenges is also a huge element of his team’s work, Loic says. “It’s all good delivering the technology, but we’re also noticing that data maturity at Chiltern is quite low,” lamented Loic. “We’re looking at different ways of solving that: bringing in e-learning from the data side of things – so e-learning in terms of how to actually manipulate data, how to ask better questions.”
In the ear of the decision-makers – and wider data sharing
Another key obstacle in the industry to harnessing data is getting the message across to decision-makers of providers in what data can do. Following a sector-wide history of under-resourcing on data in rail, Loic said that when Mary Hewitt, managing director at Chiltern, joined the business, “one of her core focusses and mantras” was around data-driven decision-making.
“If you don’t have your director mentioning ‘actually we could do econometrics on this,’ or if you don’t have someone at a high enough level saying that we would benefit from automating this data-feed, then you’re not going to get anywhere,” Loic argued.
“Mary led a whole wave of change in the way that we utilise data at Chiltern. It takes a way of being able to describe it, and I think there’s a lack of understanding of those data capabilities as a huge part of it. How do you make someone understand? Showing them. Showing people how things work. And getting them engaging to it, I think, is one of the key steps forward to doing a project like this.”
And, Loic hopes, the greater the number of digital-savvy decision-makers in providing services for passengers, the snowball effect of stacking datasets to drive business decisions will open up more cross-industry potential for providers, Network Rail, and even infrastructure firms in the private sector – to share data sources and build a comprehensive virtual landscape in rail.
“I think it’s hugely important that we do start sharing data,” Loic noted. “Let’s just talk about the passenger count project: the passenger count project is taking in data from a variety of sources, from gate-line, advanced ticket booking, CCTV; and requires those shared outputs of data,” he said.
Loic highlighted booking app Trainline as a key area where data sharing could be improved: if a user was to book a train from Leamington Spa to Marleybone, Chiltern Railways is delivering that service, regardless of way the ticket was booked. But, because Chiltern does not have a shared data agreement with Trainline, Chiltern itself misses out on a whole 25% of data from tickets purchased, meaning that Chiltern is unable to understand the experiences of one in four of its passengers, for example. “We can’t get access to any opinion, regardless of if they’re using our service to actually travel. It’s a massive, massive problem,” Loic says.
I think it’s hugely important that we do start sharing data
Last week Network Rail announced it was opening its live data source for some 1,500 lifts and 300 escalators around stations in the UK. Network Rail hopes it will allow passengers to see which lifts and escalators are working with real-time knowledge of the access points available, with Network Rail encouraging UK-based app developers to turn the data to something “really interesting” for passengers.
Though a relatively minor example of a data-centric decision-making approach, Loic sees the decision as a major positive in building that digital comprehensive view of UK rail. “I think it’s hugely important to getting a full picture of the rail industry, and, as I demonstrated with the passenger count model, being able to have the flexibility of being able to join various datasets together is important,” he said.” “I think data-sharing is definitely for me something which is not done enough in the rail industry right now, and we definitely could look into doing more with that.”