Insightful blueprints of Brussels traffic emerge in times of Covid-19

The present corona crisis has made life and work circumstances quite challenging for everybody, but for data science the lockdown restrictions created a number of very interesting real-world problems to study. The EluciDATA Lab of Sirris has fully exploited this opportunity and has managed to derive some characteristic blueprints of the traffic in Brussels.

In the context of an industrial PhD project in collaboration with Macq and VUB, the EluciDATA Lab of Sirris started in the beginning of 2020 to gather publicly available data of the traffic in Brussels, captured by Brussels Mobility at 55 locations in the city. The original aim of the PhD project, sponsored by Innoviris, is to develop an advanced trend analytics engine that facilitates understanding traffic situations in an accurate and situation-aware way.

In January, we did not have the slightest suspicion that the escalation of the corona epidemic would deliver a unique real-world dataset for our research. Herewith, some of our findings so far.

Traffic volume unevenly affected by the lockdown

As expected, the overall traffic volume in Brussels has dramatically reduced during the lockdown period. For some key locations (e.g. Keizer Karellaan, Willebroekkaai and Troontunnel), more than 50% of the traffic was retained, while other locations (e.g. Lorrainedreef, Vleurgattunnel and Stefaniatunnel exit direction centrum) have seen a substantial reduction, up to 80%, of traffic volume in comparison to the pre-lockdown period.

On average, 40% of the traffic volume was retained during the lockdown. The Brussels small ring R20 retains more volume compared to the residential areas around it. This might mean functional traffic in the city of Brussels is less reduced than recreational traffic.

Fingerprinting of weekly traffic confirms compliance with lockdown measures

When analysing the data hour by hour, day by day, we have extracted a weekly traffic intensity fingerprint for each of the monitored locations and can clearly see the measures taken one by one, e.g. the shopsand restaurants closing at the time. These weekly fingerprints are very interesting since they offer a kind of blueprint for the amount of absolutely inevitable traffic in Brussels. The clear similarity between the consecutive lockdown weeks indicates that citizens in Brussels continue to follow the measures as well as at the beginning, no relaxation can be observed.

Distilling traffic volumes associated to specific activities

The collected data spans over a time period covering three different distinct traffic situations: 1) normal referring to regular work-school weeks; 2) carnival holidays, referring to the spring half-term holidays; 3) lockdown weeks referring to the period of activity restrictions due to covid-19 measures.

Comparing the characteristic fingerprints allows to disaggregate the traffic volume into separate intensities associated with different activities. For instance, the Troontunnel shows both increases and decreases of traffic during the carnival period in comparison to the regular weeks. However, subtracting lockdown traffic volumes from the corresponding ones during the vacation week, is more intriguing. Both fingerprints summarise traffic volumes not containing school-related travel. However, the vacation week still contains a lot of recreational traffic related to school children, which is not the case for the lockdown weeks. In addition, the vacation week also captures traffic related to people who switched to teleworking or are technically unemployed during the lockdown and also non-essential activities (e.g. non-food shopping, restaurant dinners and events).

The purpose of this analysis is to illustrate that due to the unique restrictions for covid-19, we are able to split the characteristic fingerprint of normal traffic into different subcategories.

Most of the present teleworkers are typically biking to work?

Similarly as for vehicle traffic, we have extracted a weekly intensity fingerprint for biking traffic, since we also have access to bike counting data of the city of Brussels. These fingerprints show that the intensity of the biking traffic has increased in the lockdown weeks, which is not surprising, since walking and biking is practically the only allowed outdoor activity.

It is also interesting to observe that for one of the monitored locations, the Koolmijnkaai, the clear morning and evening peaks, which must be mostly representing work commuting, have fully disappeared in the lockdown weeks. This raises the question whether most of the bikers switched to teleworking during the lockdown.

Is speeding induced by too much traffic?

Another interesting measure for studying Brussels traffic is speed. The relative difference in average speed during lockdown vs. normal period was compared for 55 locations. Considering the traffic in Brussels is very saturated, it is normal to expect that for many locations the average speed has increased since the start of the Covid-19 restrictions. During a regular workday, the Troontunnel is often clogged with traffic and the maximum allowed speed of 50 km/h is hardly possible. It is interesting to observe that, although the morning and evening peaks of traffic volume in the Troontunnel have been reduced with only 21% during lockdown, the average speed has increased to the legal maximum. Thus, it seems that the modest reduction of traffic volume is already sufficient to permit fluent traffic flow.

It is also remarkable that during the lockdown weeks people seem to respect the legal speed limit more. For instance, the Georges Henri tunnel has a speed limit of 50 km/h which is rarely respected during normal times: the average speed hovers around 70 km/h during the day. However, during the lockdown weeks the average speed dropped to the maximum allowed 50 km/h. The same phenomenon can be observed during the weekends for both mentioned locations.

We believe there are two possible explanations for this phenomenon. Either, the drivers are less stressed since traffic is less hectic, there are almost no traffic jams, people are not in hurry to be somewhere in time, or the drivers who typically violate speed limits are not driving in Brussels during the lockdown. Or perhaps a combination of both?

What’s next?

The five main conclusions that could be drawn from the data analysis were only possible because of the lockdown.

The covid-19 restrictions will be phased out in the next weeks/months by the government. It will be interesting to see how fast the population will react to this and whether we will return to the original traffic patterns.

Read the full article here.