Mind the (Data) Gap
Lauren Sager Weinstein is TfL’s Chief Data Officer, working to bring value from TfL’s vast array of data by delivering data products and services, and helping Londoners journey across our network. As Chief Data Officer, she heads up TfL’s Data and Analytics department, which is a team of data scientists, data software developers, and analytics translators who provide data tools for TfL to understand customer travel behaviour and analytic tools to operate and plan London’s transport network. Lauren created TfL’s data function and has led the development of TfL’s data strategy and data capability programme across the organisation. She is now focused on leading TfL’s approach to AI.
Lauren has worked at TfL almost as long as TfL has been in existence – joining the organisation in 2002 and working across the areas of strategy, business planning, and technology development before moving into tech. Originally from Washington, DC, USA, she has degrees from Princeton University and from the Harvard Kennedy School of Government, who awarded Lauren the 2019 Alumni Award for Digital Innovation. Lauren was named the 2017 UK Chief Data Officer of the year by the CDO Club and was honoured to be included in The Female Lead’s 20 role models in Data & Technology 2017. She’s also made an appearance on the DataIQ100 List for several years.
Listen to this article
Every day Transport for London (TfL) is responsible for the safe and efficient running of over 9,000 buses, and 11 Underground tube lines. At peak times more than 540 trains are operating under the streets of London and more than 5 million passengers use the Underground each day.
Clarity of purpose
The first thing that strikes you when meeting Lauren is the total clarity of purpose. She knows exactly what she and her team are there to do: “As Chief Data Officer, I help us think strategically about how we use data better as an organisation in order to deliver services for our customers.”
That clarity comes in part from TfL’s position as both service provider and strategic overseer, and unlike the world of aviation where independent and often competing organisations try to work together, Lauren and her team often has access to the full end-to-end picture of someone’s journey. It’s a position she admits is immensely helpful: “It’s a luxury, and that allows us to think about our service holistically. Most transport organisations in the world don’t have the full end-to-end that we do.”
That luxury means she has access to a vast trove of operational data – every train movement, ever traffic light change, every tap of a contactless payment card. The trick is to know what do with it all. For Lauren, it always starts with the why: “My job is a lot of fun and it’s a lot of big questions. What does this particular piece of technology do? What business challenge and problem am I trying to solve with it? How can the data and information help make better decisions?”
It’s an approach that she believes is applicable to any industry. “You can think about those things, even when you don’t have the luxury of a really vast, expansive network like we do. How can what data you have be used to think about your processes differently?” And importantly, this isn’t a purely academic process: “It’s not just interesting. It should be because we want to take action. We want to do something, we want a policy outcome, we want to measure the effectiveness. We want to provide a service.”
It’s not just interesting. It should be because we want to take action.
And for those not blessed with end-to-end data, Lauren has some words of advice on how to encourage data sharing between sometimes competing organisations. “Everyone needs to have a confidence that the data is going to be used appropriately. You have to build trust. Step one is to build up a collective problem - a challenge to collectively solve. And you need to understand the people behind the data, what are the concerns? If it is commercial concerns, do you need to put in place protections? And then what are the techniques that make sure you are sharing what's appropriate?”
On the buses
An example of the ‘outcome first’ approach is one Lauren shares with a sense of genuine enthusiasm: “I love talking about the bus network. We have bus services all over London, and we want to provide them as efficiently as possible to provide the best cost model we can, and to attract the most customers because we collect the revenue, so planning the route network is really important.”
On the surface, running a metropolitan bus network might not appear to have much in common with the world of air traffic management or aviation, but her example is a lesson in having a very clear idea of the questions you’re trying to answer and why. “One of the things we didn’t know was, ‘where are people going?’ If you’re riding a bus in London, you tap [to make a contactless payment] when you get on, but it’s a flat fare so you don’t need to tap off again.”
TfL are using designing algorithms to better understand passenger journeys
Up until that point, understanding people’s onward destinations was a mix of guesswork and paper surveys – not an ideal way to plan one of the world’s biggest transport networks. Working out where people were heading - despite missing a vital piece of the puzzle - was a tantalising challenge for Lauren and her team. “So we said, what information do we have? What data do we have? We knew where our buses were because we’re able to geolocate them. And we also know where our customers are tapping on when boarding a bus.”
By focusing on what data did exist - and with a very clear outcome in mind - the team was able to bridge the information gap. “We built an algorithm. We depersonalised the data and then looked at the patterns of of people travelling, and then we could see where their next [ticketing] taps were – perhaps at a Tube station - and make an inference as to where they were likely to have got off the bus. We then use the aggregated pattern data every day in our operational planning and service management” What started as a data challenge based on a simple question – where are people going? - ultimately helped TfL design and deliver a better bus network, one that has helped increase connectivity across London and revenue that can be invested back into the service.
My job is a lot of fun and it’s a lot of big questions. What does this particular piece of technology do? What business challenge and problem am I trying to solve with it?
A clever guessing machine
It would be impossible to talk to Lauren and not ask her about Artificial Intelligence. With access to such a vast array of data, how keen is she to see how AI being used to help her and the team? “We’ve been using AI for years, and we mustn’t forget that, that’s fundamental for us. But generative AI is the thing that’s new and it will be transformative. It’s already transformed some things.”
But with a hint of caution, she adds: “That’s where you need to really understand what Gen AI is - a very clever guessing machine and it’s super clever because it’s got super data. Assuming the data is really good, it can get really good answers that are probably right, but you have to understand what conclusions you can draw and what conclusions are spurious.”
Gen AI is a very clever guessing machine and it’s super clever because it’s got super data
So where does she think AI most broadly it has a role to play? “We did a smart station trial where we were able to use sensors to understand what might be happening. Did someone trip and fall in a station for example? Running a safe network is fundamental to us, so could AI image detection be used for that? It’s an idea we’re looking at.”
It’s an example that mirrors the use of image detection at some airport digital tower facilities – such as Hong Kong’s use of Searidge’s digital apron management system to flag missed turnaround milestones. But in both cases, there remains a person with executive decision-making control, which Lauren sees as vital: “There needs to be someone who takes accountability for making that decision and understanding the data behind it and understanding what the data means and what it doesn’t mean. That’s always been the case, but it’s even more so with AI.”
Like TfL, Hong Kong Airport is using AI based image technology provided by Searidge in their operations centre.
Photo credit: Searidge Technologies