Would you like a commercial appointment? A consultant will call you back.
Crowd monitoring is efficient technology used to analyse the movement of people flows. Orange works with partners to offer crowd monitoring to its customers. Startup company Cropland is one of those partners. We spoke to Peter Rakers, Data Strategist at Cropland, about the applications of the technology.
“In recent years, we have noticed that companies are facing ever more significant, complex data challenges,” Peter Rakers says. “Our crowd monitoring story started with such a challenge: a few years ago, Gert Pauwels, M2M Sales & Marketing Manager of the Orange IoT department said in one of his presentations that Orange had a great deal of data, but did not always know what to do with them. I was in the audience, and I got in touch with him. We started working together, and we now offer crowd monitoring services together according to the Cropland Analytics methodology.”
So crowd monitoring allows us to follow the movements of people flows?
“Yes and no. Yes, we can follow the movements of people, but we can only do this indirectly. It is more accurate to say that we monitor the data of mobile phones. Your mobile phone will periodically contact the nearest mobile phone tower to make its presence known in order to be able to make calls and send text messages or data. When you are on the move and you have your mobile with you, your mobile will keep contacting different mobile phone towers. This proximity to mobile phone towers is what we monitor.”
So you don’t know anything about who is using a particular mobile phone then?
“No, we don’t deal with the identity of users. Orange ensures that the data are anonymous before it provides them to us, so we don’t know which SIM card is in a mobile phone, who the owner is or whether it is a man or a woman. All we see are the location data with a unique ID per mobile phone. For example, we see that a mobile phone was in Charleroi first and then in Antwerp two hours later, but that is all. The privacy of the users is fully protected. We will always restrict the monitoring of devices in space and time to a strict minimum.”
You are mainly interested in large people flows, rather than individuals then?
“That’s right: our analyses have little meaning on an individual level. If there are only a few devices in a certain area, we don’t even show them, as this could allow us to identify the users from a theoretical point of view, at least. However, when there are many devices in one place, we can analyse the crowd.”
How accurate is your localisation?
“Our analyses use grid squares of 100 by 100 m. The towers’ radiation fields go from several square kilometres in the Ardennes to just a few blocks in a densely populated urban area, but even an analysis at block level is sufficient for most events. The Antwerp 10 Miles running event is one example. You want to identify the major people flows and see when peak times occur. You don’t need localisation at street level for that.”
How is crowd monitoring applied?
“The first crowd monitoring application we implemented together with Orange was an analysis of the Antwerp Sinksenfoor fairground in 2015. The Sinksenfoor fairground had just changed locations to an open space where visitors could enter and leave wherever they wanted. This made it difficult to count the number of visitors with security gates. We developed a proof-of-concept for crowd monitoring with Orange, and the results were promising. A few more test cases then ultimately led to a one-year city monitoring contract in Antwerp for which we measure 20 places in the city permanently, and we monitor major events. The contract with the city of Antwerp has meanwhile been extended by two years. Based on this experience, Orange developed a standardised report - the City Monitor Report - which allows the collection of interesting data in towns and cities.”
So crowd monitoring provides cities with lots of interesting information.
“Yes, it does, in three different areas: safety, mobility and city (or event) marketing. As far as safety is concerned, we can monitor people flows anywhere in Belgium almost in real time and provide feedback to a disaster coordinator. At major events, such as the New Year's fireworks or the 10 Miles of Antwerp, such information is invaluable. Crowd monitoring also gives us unique insights in terms of mobility. For example, we used this technique on behalf of the Public Service of Wallonia to analyse cut-through traffic after the introduction of road tolls for lorries and the traffic congestion around the Pairi Daiza animal park. The latter project even went on to win the Agoria Smart City Award Digital 2017.”
And how does crowd monitoring benefit city marketing?
“Bringing events such as the Tour de France, the Tour of Flanders and the huge mechanical marionettes of French theatre company Royal de Luxe to the city of Antwerp involves significant costs. Of course, the city considers such events as investments and hopes the local services and tourism industry will benefit. Crowd monitoring contributes to a well-founded cost-benefit analysis. It can show you how many foreign tourists are visiting the city and how long they are staying, for example.”
All this is based on the signals that mobile phones send to the towers in their area. What are the key advantages of this type of crowd monitoring compared with people counters and cameras?
“The main advantage is that crowd monitoring can be applied anywhere in Belgium without investment. The technology can be used wherever the Orange network has coverage. Another advantage is that it allows us to look back in time, as Orange stores its location data for a year. When the Public Service of Wallonia approached us for an analysis of cut-through traffic, we were able to compare the location data before and after 1 April 2016, when a toll was introduced for lorries. The counts for this project are still on-going.”
Are there any challenges in terms of crowd monitoring?
“There are some major challenges, certainly. The most important ones are the extrapolation factors, which are different in every situation. I mentioned earlier that we monitor mobile phones rather than people, but our customers are obviously interested in people rather than mobile phones, so we still need to convert the number of mobile phones to the number of people. And that depends on the situation, of course. When kids visit a big show with their parents, the number of persons per mobile phone will be different than when we analyse the Tour de France. We first visit the target group for every event we analyse with crowd monitoring. We also ask questions to the people on site to fine tune our extrapolation. The interaction with the customer is crucial here, because it allows us to make optimal use of the knowledge with regard to the event itself.”
The raw data alone don’t tell us much then. You also need domain expertise to interpret the data?
“That is correct. You should always look at the baseline activity. The Antwerp 10 Miles running event is a striking example. Both the start and the finish are located between big apartment buildings on the city’s left bank. A lot of people gather in a small area, but if you want to measure the number of people visiting the 10 Miles, you need to deduct the number of persons living and working there every day, of course. Crowd monitoring requires so much more than just counting.”