Data can be a powerful tool in urban mobility, but what happens when numbers don’t tell the whole story? As city planners and citizens embrace big data and new technologies like artificial intelligence, various risks can arise, affecting crucial decisions about mobility and resulting in serious consequences for how we move around. All the more reason to avoid the common missteps and overlooked strategies that can lead to data mismanagement in the first place.
To help us navigate the nuances of the numbers, UMX spoke with Xavier Tackoen, head of Espaces Mobilities in Brussels, where he develops strategic mobility projects. (You may have seen him in several UMX videos on autonomous vehicles!) He’s also the co-author of our free online course The power of Mobility Data: Discover how every move matters, which is based on a series of hands-on Mobility Masterclasses that Espaces runs throughout the year with EIT Urban Mobility. Drawing from his extensive experience, Xavier provided us with valuable advice and actionable recommendations to manage mobility data effectively.
Data is everywhere, but mostly out of reach.
Cities have been talking about mobility data for a decade, but only in the last couple of years has the volume grown exponentially to what we encounter today. The course goes through all the different ways that humans, vehicles and city infrastructure are connected to surround us by data. But it begs the question: Who’s responsible for the generation and collection of so much new data? The private sector. “They’ve been able to integrate and connect many devices,” as Xavier explained, “and bring new insights and value to what we do in terms of transport planning.”
The word “private” also means there’s a lot of red tape around that data. Many believe that it’s far easier to retrieve data on, for example, the number of e-scooters in a city, than it is because of restrictions placed on many Big Tech companies. Xavier noted that people “underestimate the impact of GDPR on the sample size and what we can really see.” Then there’s a dilemma between privacy plights and transport planning. He continued, “It’s a tricky balance between protecting citizens and mobility optimization, and the EU is taking the lead in that field,” with measures like data protection procedures and the more recent AI Act.
Confusion over what to measure and how to interpret it.
While large and juicy data sets from the private sector can look awfully appealing, using them is contingent upon understanding what you’re using the data for. Xavier pointed out, “If you don’t know what you’re looking for then it’s better not to buy any data.” On the flip side, it’s dangerous to be too confident about what you’re looking for. Some decision-makers go into the data analysis process with preconceived notions of what they expect to see and, as Xavier added, “sometimes they don’t want to be contradicted by the data we share.”
In addition, it’s common to buy data sets for a study but forget to measure the baseline. “You need to gather data before, during, after, and so on,” Xavier described. “Data should be a package.” It’s also critical to talk about relative figures instead of real ones. For example, Xavier mentioned a case (covered in more depth in the course) regarding a highly anticipated new tram line in a main province outside of Brussels. As it turned out, 87% of all trips in that province stayed local and only 13% would use that line. As a result, Xavier said, “Even if half of those remaining people take the tram to go to Brussels in the future (compared to a third now), the overall outcome will not change significantly.” In other words, never lose sight of the scope of your data within the greater context.
Steering clear of conflicting and biased data sets.
When purchasing data sets, public institutions tend to take a tender route in which companies compete over factors like quality, price and more to be the singular data provider. However, as you can probably imagine, several things can go awry in the data generation and collection process. But with just one source of data, what if it’s wrong? That’s why Xavier recommends purchasing data from at least two providers when possible. “If they both show the same thing, then you are convinced that this is reality,” he explained. “Otherwise, my fellow consultants and I will compare the data with traditional data, like household surveys or ticketing metrics to find out if we see the same trends.”
Moreover, Xavier warned about a “black box” effect from providers to avoid data bias and misinterpretation. He brought up examples of underlying trends such as gender effects on mobility, as well as outliers that can skew the data, such as weather, strikes or even a Taylor Swift concert. “Often people don’t check in advance what happened in the area before buying or analysing the data,” he emphasised. So while big data is valuable, it’s not meant to replace all other forms of data. “It’s quantitative data that compliments qualitative data. You can’t really merge them,” he continued. “You show the two sides of the coin” to guarantee more precise insights.
Artificial intelligence, data management and best practices.
We would be remiss if we didn’t mention how artificial intelligence adds another layer to this story. Xavier was hopeful that AI will accelerate the data analysis process: “Whereas 10 years ago you had to go into Excel files or databases to extract data on something like how many public transport lines were delayed last month, now I can prompt AI to do that.” But he also remained cautious against a so-called “prompt generation society” because AI can exaggerate existing biases. “If the prompt is not written correctly or is not understood correctly by the large language model (LLM),” he advised, “then you might get completely wrong results while giving the impression that the results are accurate.”
On a similar note concerning multiple data sources, Xavier encouraged that data be overseen by the proper administrative body. To avoid cities, municipalities and public transport operators from replicating or complicating work that’s already been done, he suggested that there be a “common methodology and common vision on data at a high national level.” By way of illustration, he brought up congestion since there’s no universal way of calculating it: “Does congestion mean waiting at the traffic light for 10 seconds, 10 minutes, 10 hours? Is it in Cairo, Brussels or San Francisco?” Furthermore, providers like Google Maps might not be delivering the same data as TomTom, which is why good data management is essential to developing public policy and partnerships that cities need to set up with private providers, especially as shared mobility and autonomous vehicles continue to pop up.
If that seems like a lot to take in, you’re right… So start small! Xavier advised that cities get the ball rolling with specific, in-depth cases to “convince both themselves and their stakeholders of the data’s value by showing the hypothesis and all of the steps, as opposed to a nice graph.” (The final module of the course outlines his 10-step strategy.) Ultimately, transparency is the key to effectively understanding and using our mobility data. As Xavier stressed, not only does it prevent data from being misconstrued, but it also holds experts accountable and can even garner more trust among citizens. “People can understand that things are not perfect, which is why we always want to show as much as we can,” Xavier said. “They’re not against decisions; they’re against blind decisions.” With these strategies in place, we get to the bottom of what the numbers are trying to tell us, as complex and imperfect as it might be.
Adina Rose Levin
Adina Levin was born and raised in Chicago, and clocked in over 10 years in New York City before moving to Barcelona. As a freelance writer and creative strategist, she explores cities, culture, media and tech.
Xavier Tackoen
Xavier Tackoen earned an MBA from ICHEC (2002) and a Transport Management Master's from CIEM (2003). With a decade in academia, he's been with Espaces-Mobilités since 2009, specialising in transport strategies, new mobility services, and public space design, focusing on MaaS, shared, connected, and autonomous mobility.
Login with Google