In April 2020, the European Commission (EC) asked European Mobile Network Operators (MNOs) to share fully anonymised and aggregated mobility data in order to support the fight against COVID-19 (European Commission, 2020a,European Commission, 2020b) with data driven evidence.
The value of mobile positioning personal data to describe human mobility has been explored (Csáji et al., 2013) and its potential in epidemiology studies demonstrated (Wesolowski et al., 2012, Jia et al., 2020,WU et al., 2020,Kraemer et al., 2020) in literature.
The new initiative between the Commission and the European MNOs relies on the effectiveness of using fully anonymised and aggregated mobile positioning data in compliance with ‘Guidelines on the use of location data and contact tracing tools in the context of the COVID-19 outbreak’ by the European Data Protection Board (EDPB, 04/2020).
This work introduces an innovative way to map natural human mobility through fully anonymised and aggregated mobile data. Maps showing natural mobility are based on the usual patterns of citizens’ mobility and can be compared with maps of administrative areas.
Indeed, the mapping of human mobility patterns has a long tradition in settlement geography, urban planning and policy making. The idea behind mobility patterns is the identification of a network of aggregated inbound and outbound movements across spatial structures for a given time scale (for example, daily, intra-weekly, seasonally, etc) according to the scopes of their use. These patterns have been called in several ways; the followings list includes just a few variants of the same concept:
— ‘commuting regions’: the identification of relatively closed regions of daily moves of residing population based on commuting data from censuses (Casado-Díaz, 2000,
Van der Laan, 1998).
— ‘functional regions’: a tool used to target areas of specific national and European policies (OECD, 2002). There are several natural areas of application of functional regions including employment and transportation policies, environmentally sustainable spatial forms, reforms of administrative regions, strategic level of urban and regional planning and a wide range of geographical analyses (migration, regionalisation, settlement system hierarchisation) (Andersen, 2002,Ball, 1980,Casado-Díaz, 2000,Van der Laan, 1998).
— ‘functional urban areas’: cities with their commuting zone (Eurostat, 2106, Dijkstra et al., 2019). They are generally identified by a densely inhabited city, together with a less densely populated commuting zone whose labour market is highly integrated with that of the city.
— ‘overlapping functional regions’ (Killer and Axhausen, 2010).
The most common data sources for the above-mentioned studies are by far the population censuses and ad hoc pilot surveys.
This study proposes an alternative method to define highly-interconnected spatial regions (i.e., forming dense sub-networks); only fully-anonymised and aggregated mobility data are used to this end. The data-driven regions identified through the proposed method are referred to as ‘Mobility Functional Areas’ (MFA).
Although mobile data has been used in the past in a pilot-study on mobility in Estonia (Novak et al., 2013), the present study adopts a new technique to define mobility functional areas (MFA), which is based only on aggregated data, and extends the research to 15 European countries (14 member states: Austria, Belgium, Bulgaria, Czechia, Denmark, Estonia, Spain, Finland, France, Greece, Croatia, Italy, Sweden, Slovenia, plus Norway).
In a policy making perspective, especially related to the COVID-19 pandemic, the insights resulting from this analysis may help governments and authorities at various levels:
a) to limit all non-essential movements across specific geographic areas, especially in the initial phase of a future outbreak of the virus, to limit spread while also limiting the economic impact of such measures outside the MFA;
b) to apply different physical distancing policies in different areas, according to their specific epidemiological situation.
In the absence of any other information, most of the governments are forced to use administrative areas, such as regions, provinces and municipalities to impose physical distancing measures and mobility restrictions. Nevertheless, administrative boundaries are static and do not reflect actual mobility. On the other hand, both the potential spreading of the virus and the territorial economy strongly depend on local mobility (Iacus et al., 2020).
Although these aspects cannot be taken into account in this work, the hypothesis is that the implementation of different physical distancing strategies (such as school closures or other human mobility limitations) based on MFA instead of administrative borders might lead to a better balance between the expected positive effect on public health and the negative socio-economic fallout for the country. Despite the evident potential benefits, it must be noted that while administrative areas (hard boundaries) are well recognised by citizens and make it easy for the administrations to implement physical distancing and mobility restrictions, further coordination efforts would be needed to apply such limitations based on MFAs.
This work is organised as follows. Section 2 describes the data sources used in the analysis. Section 3 explains in details the concept of MFAs and, along with Section 4, describe the methodological approach to identify MFA pre- and post- lockdown measures; the evolution in time of the MFAs is presented through a case study for Spain. Section 5 is a quick review of the results for each of the remaining 14 countries considered and finally Section 6 shows an overall view of the MFAs across Europe (15 countries analysed).