The COVID-19 pandemics is still impacting many countries around the world. Nations where the first wave of the pandemics is over are now facing the challenges of reopening phases. After many facilities, industries and schools have been closed, the question is now what to reopen first, which locations are safer and how to avoid overcrowding. Measures, such as smart-working and on-line courses for schools, are being discussed by many governments in Europe. Never as of today, the critical role of public transport and transportation systems, in general, is evident to the public opinion.
The shift in individual behaviour is another challenge that will have to be addressed and hard to foresee to the present day. Any planning related to transportations and activities reopening could fail because of unpredictable changes in what activities individuals typically do, where do they typically travel, why they move there and so on.
In this report, we show an update to what we presented in a previous post. There, we showed how it is possible to understand the preference of individuals in terms of visited locations. We performed that activity by merging open data about Points of Interest coming from OpenStreetMap, and High-Precision Anonymized Location-Based data (HFLB) provided in the context of CUEBIQ’s Data for Good Program.
In the following, we will present some interactive charts that allow exploring how individuals residing in different Italian regions visited different locations before during and after the lockdown phase. It is interesting to notice that while the country has been unequally affected by the pandemics, the effects of the lockdown have been symmetric all over the country, meaning that the individual response has been independent of the immediate risk of contagion.
The methodology used is the same as the one adopted in the previous post. We start using the High-Precision Anonymized Location-Based (HFLB) data collected via mobile phones app, and anonymized at the moment of their collection. These data have been provided in the context of CUEBIQ’s Data for Good Program. This first-party data is collected from anonymized users who have opted-in to provide access to their location data anonymously, through a GDPR-compliant framework. We then join these data with the ones fetched from OpenStreetMap, specifically the locations and categories of the Points of Interests (POIs) found in a given area, so as to quantify how many individuals have visited locations such as:
Workplaces and Homes where identified with a certain degree of approximation, but we cross-checked with Italian Census data to see if our guesses were correct. Note that the intrinsic noise of the position we had allows us to assign a location with a precision of around 100m.
By dividing individuals according to the region they live in, it has been possible to visualize the evolution of their visits to the different locations categories in time, from the beginning of January 2020 to the end of June 2020. The time frame covered by the analysis goes from a couple of months preceding the national lockdown, up to some weeks after the relaxation of the latest restrictions. The following figure allows to explore the time spent, the amount of users and the amount of stops observed in each kind of location.
Instructions for Fig. 1
By visualizing the number of individuals that visited each location in time, it is possible to have a better understanding of how different areas reacted to the restrictions and the reopening phase. The interactive map below allows for the exploration of how the frequency of visit for the various locations categories evolved in time in different regions. While in many locations the number of visits went back to the pre-COVID levels, some of them are still less visited than the pre-COVID. This outcome is partly due to the restrictions that are still in place (univerisities are still closed at the time of writing) and partially stems from the fact that many activities decided not to reopen yet, either for safety reasons or economic convenience.
Instructions for Fig. 2
As we explained in our previous post, the characterization of how individual adapted their habits in response to the COVID-19 outbreak is of utmost importance for the planning of the reopening phases and for the organization of the society during what someone calls the New Normal. New habits will affect the way in which cities will evolve in the future, and any form of planning -from transportation, to education and work- will have to take them into account. Data-driven approaches (i.e., Data-driven decision making) will play a key role during the design of the policies to be implemented and in the monitoring of their efficiency. This is particularly true because of the short time needed to measure the reaction of people to such policies and the level of details that can be achieved when using such large data bases. In this spirit, we already shown a paradigmatic example on how to inform policy-makers during the reopening phase: the modelling and forecast of the Social Distancing Index (SDI) on the public transport in Rome, that we presented in this post. We will further expand that model in collaboration with “Roma Servizi per la Mobilità” and report our findings in a future post in which we will explain how the information reported here will be used to plan the reopening of schools in the Metropolitan Area of Rome. So, stay Tuned!.