Crowd monitoring in cities gained considerable attention with the outbreak of the COVID-19 pandemic. The City of Ghent, however, was a step ahead, as they were already working with crowd data well before the start of the pandemic in 2020. To understand how and why cities should strategically look at crowd data, we interviewed Justine Ottevaere from the City of Ghent’s Data and Information Service, to learn more about her experience developing a crowd monitoring tool.

JO. It started in 2019 with the Smart Flanders Working Group, when 13 of Flanders’ cities came together to tackle common challenges on open data and GDPR. At that time, data suppliers – the telecom companies for example – were largely in control of what data was available and which questions could be asked. And there was no transparency about their technology and any changes in their methodology over time. Together, we wanted to create a definition book (shared profiles of visitors in our cities, shared meanings for the data elements we wanted to collect and analyse) that would facilitate collaboration and make the voice of our cities stronger.

Once the challenge in the Smart Flanders Working group was clear, I took the work back to Ghent to collect our own definitions, data and experiences. In Ghent, I found that the different city services were buying data on their own, sometimes common data from the same suppliers, or even the same data from different suppliers. By coming together, we could share data, semantics, and definitions. For example, both the Ghent Tourist Office and Economic and Entrepreneurial Support Service would be interested in data on people coming into the city for shopping.  So, we bought the data together – meaning that the insights from that data became a stronger story that captured the whole city.

What started as an informal working group in Ghent became a fixed structure known as the drukteclubje (crowd-club), with about 10 regular members. Now we meet every month to share needs, experiences, learnings, data and insights on that data and our projects. Other department representatives, experts and project managers also join from time to time to ask questions or share their experiences.

JO. The drukteclubje started with domain experts from Ghent Tourist Office, the Festivities and Ambulatory Trade Service, the Economic and Entrepreneurial Support Service, the Mobility Department and the Innovation Cell Foresight. Combining the teams allowed us to combine data, share insights and conduct data analysis. Since the outbreak of COVID-19, the Environmental and Climate Service has also been involved as parks have become busier and they need data on the number of visitors to keep people safe and to preserve nature and animals.

To promote the drukteclubje, we have an internal platform to highlight the short and long-term work we are doing. We also have internal events – called ‘sandwiches’ where we present several topics and afterwards we share ideas and discuss the topics while eating sandwiches.

The benefits are that we share experiences and insights which saves time and money. We have created an open work environment where we are able to experiment and to fail, as long as we learn something from it. Now the challenge is to translate our experiments into long-term activities.

JO. Sharing data means that the budget is spent more efficiently. It’s also a new way of working. When I first joined Ghent, I was amazed at how little sharing occurred – every department worked alone, without cooperating on projects. So, it’s a big positive change now to be working together with different departments in different cities. Sometimes, it’s not easy because everyone has their own budget and own rules – and some department leads are more open to joining than others. In a big organisation such as the City of Ghent, it is sometimes hard to find the time and money to experiment because the short-term value is not always clear from the beginning. So, it is important to highlight the quick wins and translate them into valuable long-term wins. Generally, though, people have seen the benefits of our work.

JO. In the beginning, our working group met less frequently, but COVID-19 was actually a good thing for our work! Things began to move quickly, and budget was made available. There was growing support from different departments and from politicians who saw the need for real-time data and understood that it is important to have a good relationship with data suppliers. For example, when the shops were due to open in two weeks’ time in mid-December, we needed access to real-time data very quickly to guarantee safety and minimise crowding in the streets. Since we had a good relationship with telecom providers, we could get valuable real-time data from our shopping streets and there was a budget to make developers available as well. Technical issues that we had in our roadmap for later became, suddenly, very relevant – and now our technical team at District 09 are working on real-time APIs.

JO. The need for real-time data has reduced a bit – it’s too expensive and not always necessary. Although we now know how to access real-time data should there be a need in the future. We need long term trends and numbers, not real-time data (at this moment). We had our quick wins and learnings thanks to these urgent matters and now we have a roadmap for the future. The work we are doing on combining data and on APIs is applicable for air quality, liveability, the housing market, water levels etc. By combining all this data, we can learn more and make more predictions. So, it’s nice that we’ve had the chance to learn and experiment. Now we can prepare everything and take it to the next level in the future.

JO. Crowd monitoring is very sensitive – politicians are aware of the privacy issues, so to ensure the support of citizens, we need to be transparent. You can’t just put-up sensors and cameras all over the city for the sake of data. When you place them, you need to think carefully about how you share the message of your work. Who gets to see the insights, how do you manage the data, and how long is it stored? During the VLOED project we want to get predictions and analysis of the use of shopping centres in Ghent and Bruges. There, we will be working with a physical living lab in the shopping centre. In the lab, we want to inform the people that they are being counted as a part of the crowd, but that we are not interested in individual data – rather, we want to count them to understand our city better and keep it liveable and safe.

JO. The most important takeaway is that we need qualitative long-term data – experimenting with small bits of data is fine, but to see valuable predictions, trends, evolutions, a baseline of good data is necessary. For example, gathering data over many years in order to collect daily information on who visits the town, where they park their car or bike, how they enter or leave the city, when are the peak times at train stations, etc. We need this long-term data. We also know that data costs money. However, expensive data isn’t the answer to everything, you don’t know all the answers just because you purchase data. We understand that balance now. Also valuable was the ability to experiment. We learnt that you don’t need the complete picture right away.

At the policy level, it was received very well. This gave us the opportunity to explore both business and technical aspects of crowd data. We have better access to long-term data, and a good relationship with the data suppliers. The mayor was involved in the project during the COVID-19 crisis, so we have had the chance to bring our work on data to a higher level and ensure the need and capabilities are understood.

JO. I learned a lot during the developer week sprint in Amsterdam in July 2019. I met people working on different applications – monitoring the city in terms of where to be and where to park the car. It was insightful and helped me create an image for Ghent as well. We also learnt what the need and challenges are elsewhere – for example, in Bradford – as well as working on our own tool for Ghent. Now we are working with Bradford and Dordrecht on a ‘crowd flow dashboard’ – which we see more as a ‘data dialog tool’ to bring developers, data engineers, analysts, and domain experts together on data. That data can be about crowd monitoring, but other use cases such as air quality and flooding are also useful.

JO. Right now, we are collaborating with the Department of Urban Development which is reorganising the Overpoortstraat. It is the favourite neighbourhood for our students to party in, but it needs to be liveable for residents too. We want to monitor the area before, during, and after the changes. We will also monitor the crowd during the city’s Lichtfestival – a festival of light with art installations on a 7km walking route. In the long term, we want more stable data, more long-term data, and thereby to build public and political support by being transparent and open. We also want to develop an analytic tool that starts from crowd monitoring but can also cover other issues from different city services.

JO. It’s a broad and fruitful theme that helps a lot of city departments and services. Crowd data can improve liveability, safety, and the economic situation for inhabitants, visitors, workers and students. It’s also an innovative way of working, creating cross-department collaboration and bringing experts and developers together.

There is also potential for more insights if we compare cities across Europe. Imagine comparing the patterns of a tourist or a shopper in Ghent with Bruges or Antwerp, for example. To do this, we need to agree on definitions to compare data across cities. We have this at a Flemish level, but it would be nice to expand on those semantics to share insights with Bradford, Amsterdam, and other cities all over Europe.

At Bax & Company, we’re lucky enough to work with experts in emerging and existing fields from around the world. Beyond Bax is an opportunity for us to share some of their knowledge with you.

If you have any suggestions or further questions, please do reach out to a member of our team: