CSIC uses mobile phone data to study efficacy of lockdown on spread of COVID-19

A new CSIC [National Scientific Research Council] project uses computer science and data science techniques to observe how the lockdown measures taken to halt the spread of the disease COVID-19 are proving effective. The results will be key to improving social distancing strategies taken in future outbreaks of this disease and of others. In order to carry out this research, a multi-disciplinary team with experts in computer science, demographics, physics and movement studies are analysing anonymous and high resolution big data obtained from telephone operators and map servers. These data explain how mobility patterns and social contact have changed since the start of the lockdown.

The project, pre-financed by the CSIC, thanks to the donation received from AENA, is coordinated by the scientists José Javier Ramasco, from the Institute of Complex Physics System (Spanish acronym: IFISC, a joint CSIC and University of Balearic Islands centre) and Frederic Bartumeus, from the Blanes Advanced Studies Centre (Spanish acronym: CEAB-CSIC) and the CREAF [Centre for Ecological Research and Forestry Applications]. It also involves the participation of teams from the Institute for Economy, Geography and Demography (Spanish acronym: IEGD-CSIC), from the Institute for Physics of Cantabria (Spanish acronym: IFCA-CSIC), from the National Biotechnology Centre (Spanish acronym: CNB-CSIC), as well as scientists from Pompeu Fabra University and the National Epidemiology Centre- Carlos III Health Institute (ISCIII).

How to lift the lockdown and when

Once all the data is gathered, the team simulates different scenarios and strategies for social distancing and helps with decision-making. The results are key both for deciding whether a stricter lockdown should be activated and to plan for the safe and effective lifting of the lockdown. "We hope that the results serve to better understand the effects of the lockdown on the spread of the disease, but also help in decision-making related to the lifting of the measures, to see whether or not it is better to end the lockdown gradually", explains Frederic Bartumeus.

"To achieve this goal the project includes several phases that are being carried out in parallel", explains José Javier Ramasco. "Firstly, mobility is characterised, which is being coordinated by the IFISC based on the contribution from various data platforms: information, for example, from online social media and mobility patterns captured from mobile phone records. In this latter case, the data are collected by the operators and companies that are taking part in the project, which provide the research team with aggregated travel flows between different areas" specifies the researcher. In no case is individual information accessed.

A second aspect is the change in conduct of people due to the perception of risk. The CEAB and IEGD are carrying out surveys and implementing mobile phone applications to quantify these changes, trying to estimate the adherence to personal protection measures by people and the changes in the amount and quality of personal contact. "This information is crucial for understanding the contagion process", indicates José Javier Ramasco.

Lastly, all these data form part of the computational models being developed by the IFISC and IFCA to study the different scenarios to exit the crisis. "The lockdown has been widespread and relatively sudden, but to avoid new outbreaks it is necessary to use simulators capable of assessing scenarios with different rhythms to return to normality, both by sector and by geographic area", warns José Javier Ramasco.

Epidemiology in the future

The project uses artificial intelligence tools and data science, and integrates big data in real time on human mobility, geo-localised surveys and computational models. This is a new way of undertaking epidemiology which combines computational epidemiology, digital demography and human mobility models. "The study will take into account such important aspects as the spatial distribution of the population, their age structure, and the distribution and characteristics of social health centres (hospitals, local health centres, and care homes for the elderly). We can see how the contention measures have changed the mobility and conduct of people", comments José Javier Ramasco.

The information and models to be developed during this research study will be made available to the public for their future use following an open data model under FAIR (Findable, Accessible, Interoperable, Reusable) principles.

A second long-term goal is to establish the basis for a computational epidemiology network in Spain, as in other countries, and a series of interoperable analytical tools based on epidemiological theories, data science and artificial intelligence, to report decisions to be taken in future situations of epidemiological crisis which, as the scientists say, is something that "has already happened on several occasions since 2009 and is likely to be recurrent in our globalised and interconnected world".

Non official translation