How our mobility data can help predict and prevent the spread of COVID-19

“Blue is definitely my favorite color these days,” says Kenth Engø-Monsen.

The senior researcher at Telenor Research gazes at the blue dots popping up on his screen. The darker the shade of blue, the more Engø-Monsen smiles.

“What we have here is data on people’s movement from one specific area to another. A red color means we have an increase in movement, while blue signals a decrease. The shades indicate the frequency. The darker it is, the lower the frequency,” Engø-Monsen explains.

In front of the researcher is a digital map of Norway covered in blue. Much of it dark blue.

“Since the Norwegian government announced its lockdown measures on 12 March, we have seen exceptional changes in mobility. In some places, the inter-municipality movement has dropped by 65 percent. We know this because the data outlines a detailed view of the population’s overall travel patterns, based on mobile signals to base stations. The statistics are fascinating, but more importantly, this information can help health authorities in their work to prevent the spread of COVID-19,” he says.

Since January, before any cases of COVID-19 were detected in Norway, Telenor had already started to provide the Norwegian Institute of Public Health (NIPH) with mobility data on people’s movement between the country’s 356 municipalities.

“Knowledge about a population’s travel pattern is vital to understanding how an epidemic spreads throughout a country and thus the population,” says Engø-Monsen.

At NIPH, a COVID-19 taskforce uses Telenor’s mobility dataset to predict plausible spread scenarios of the virus in Norway.

“We have created a metapopulation COVID-19 transmission model. The model consists of three layers: a population structure of each municipality; a disease transmission model of each municipality; and the mobility data describing movements between municipalities. From it, we can estimate the current and future trajectory of the epidemic down to the municipality level,” says Birgitte Freiesleben de Blasio, Department Director at NIPH.

de Blasio says the information is used to quantify the number of hospitalisations and ICU beds needed, both at the local and national level.

“The model provides situational awareness and predicts the course of the epidemic, as it unfolds. During public health emergencies, like the current COVID-19 pandemic, public health authorities need to make decisions on interventions measures and target response needs. In these situations, mathematical models are valuable tools for preparedness planning and decision making,” says de Blasio, adding that the data also let health authorities examine if government actions such as school closures, tele-working and banning cabin trips have reduced the spread of the infection, and impacted the population mobility.

The generated data NIPH receives from Telenor is based on phone signalling.

“To be able to have good coverage, your phone will always connect to the closest possible base station, and through these base station connections, you will leave behind a location trace. The data based on these location traces are extracted from Telenor Norway’s more than 8,100 base stations located all across the country. We now count this aggregated people movement every six hours, every day, in order to give NIPH access to the most updated and comprehensive datasets available on Norwegians travel patterns,” says Engø-Monsen.

In the process of gathering data to NIPH, his team at Telenor research is working closely with the Mobility Analytics team in Telenor Norway.

“About 80 percent of all the data traffic in Norway passes through our network. It is important to note that this dataset gives insight in the whole population, not only Telenor’s market share,” says Marte Ruud Sandberg, Product Marketing Manager in Telenor Business IoT and Big Data.

When handling user data, questions regarding anonymity often arise. Engø-Monsen ensures that no one runs the risk of being exposed.

“We anonymise all data, so it is not possible for anyone to identify users. Our goal is not to know where a single individual travels, but to get an overview of the general travel pattern in the different municipalities,” says Engø-Monsen.

The researcher explains that every mobile operator is constantly collecting data about subscribers’ locations in order to provide the most optimised service possible.

“When someone is calling you, the system needs to know what cell tower your handset is connected to so that the call can be directed to that cell tower. This gives an approximate device location, but all this big data is only flowing through our systems for a short period before it is deleted,” he explains.