Skip to main content

Researchers from ÖK and ELTE use public iPhone mobile tracking data to study the effects of spring COVID measures in nine European countries


Researchers at ELKH Centre for Ecological Research (ÖK) and Eötvös Loránd University (ELTE), in collaboration with their Australian colleagues, have examined the relationship between the spring COVID measures and people's behavior based on public mobile phone mobility data in relation to mortality rates in nine European countries – Belgium, the United Kingdom, France, the Netherlands, Germany, Italy, Spain, Switzerland and Sweden. The research study has established that the timing and severity of lockdowns, as well as the cooperation of people and the observance of social distancing, had a crucial influence on the number of deaths. The fast and clear instruction to 'Stay Home!' saved many lives where it was taken seriously by the people. The research study was published in Scientific Reports on 18 January 2021.

After a new type of coronavirus spreading from China had led to the global epidemic in March 2020, in the absence of vaccines or effective drug treatment, countries around the world had only one tool to use: to lock down or limit their borders, their economies and the daily lives of the general population. Given that there had been no pandemic of this magnitude in the last 100 years, there was no knowledge available on the impact that the drastic measures that had to be introduced would have on the mobility of people and subsequently on the evolution of the pandemic. However, evaluating the effectiveness of lockdowns could be vital, as responsible decisions can only be made in the future based on the findings of this evaluation.

Researchers have found that Apple’s publicly released mobile phone mobility data are a surprisingly good predictor of the extent to which face-to-face encounters would decline. The reliable mobility data made available in this way were compared by the researchers with the dates of official lockdowns and the shape of mortality curves. (Case numbers were not used as an indicator because they are highly dependent on the amount and methodology of testing, whereas mortality data are much more reliable.)

It is clear from the mobility data that the mobility of iPhone owners began to fall around the lockdown and then settled to a much lower level. As the number of iPhone owners in Europe is quite significant, these data can effectively characterize the mobility of the whole society.

Daily mortality and mobility data for the nine European countries studied (a – i). Time 0 corresponds to the day when a country first reported ≥ 5 deaths per day. Top: the red and blue lines indicate the growth and decay phases. Vertical line: national lockdown date. Bottom: mobile phone tracking data, normalized to pre-epidemic averages (M1).

“We examined whether there was any correlation between the number of deaths and the official start of the lockdown and the actual lockdown affecting the people. The delay in lockdown was calculated based on the time that elapsed between the date of death of the fifth patient who died in the country and the date of the lockdown. It turned out that there is no strong correlation between the official lockdown date and the number of all fatalities in the spring wave," said István Scheuring, one of the authors of the study, a member of ELKH's Evolutionary Systems Research Group at the Centre for Ecological Research and of the MTA-ELTE Theoretical Biology and Evolutionary Ecology Research Group. “At the same time, we realised that the number of deaths – and the number of those infected – strongly correlates with the severity of the lockdown. This, in turn, can be estimated on the basis of how much people's mobility decreased compared to the usual rate of mobility.”

The mobility data of iPhones clearly shows the differences between the lockdowns in each country. The two endpoints were Italy and Spain (where drastic measures were taken) and Sweden (where people’s lives were nowhere near as severely restricted). Countries also differed in how quickly people’s mobility declined after the start of the pandemic. The onset of the pandemic also varied by state, presumably influenced by the structure of local communities, the mobility of local people and social life. While in Italy, for example, in many places several generations of families live together, in Sweden 40 percent of households have only one person.

“The success of the lockdowns depended on three things: how quickly the pandemic had spread before the lockdown, how quickly the lockdown was introduced by the authorities, and how strict it was. All this can be clearly seen in the mobile phone tracking data, so this information will be used in the future to check the effectiveness of measures,” István Scheuring argued.

The results show that after the peak of the pandemic, the more people stayed at home, the faster the number of deaths began to decrease. In addition, the pandemic peak occured much closer to the date of the lockdown in countries where people took the need to stay home more seriously. The number of people who died at the time of the pandemic peak was higher in countries that were late in introducing the lockdown.

“The speed of the decision on lockdown is very important. The growth of a pandemic is exponential, and exponential processes are insidious: the initial, seemingly minor growth turns into a steep rise from one moment to the next, so a delay of a few days results in more lost lives,” István Scheuring summed up the conclusions of the research.

The whole article can be found here: