The term mobility refers to a kaleidoscope of meanings, all of which have to do with the idea of change and the concept of opportunity. Far from being just a displacement from a place or position A to B, it elicits a panoply of opportunities for education, work and leisure. Being able to move freely enriches our human experience and the potential of our communities and humankind as a whole. A population that can move is more culturally, economically and socially aware. The COVID-19 outbreak has impacted on a mobility context far from being optimal both in urban and in peripheral areas, triggering possible long-term consequences. The Pandemic has led to exploring a “new normal”, which is also composed of the new restrictions that people experience in terms of mobility and social contacts in general.
However, these unprecedented challenges could provide the opportunity to transform the existing mobility ecosystem and our habits, promoting a more resilient society, based on enhanced accessibility, social inclusion and well-being for all. In total antithesis with the current lifestyle, marked by the consumerist philosophy of disposable, the Pandemic could bring about a complete reversion of materialism. Quoting the ancient Japanese art of Kintsugi (金継ぎ), through which broken objects revive, our societies could experience resilience and rebirth.
It is then of paramount importance to explore the new options that only an extreme event such as the COVID outbreak could make visible with a transformative eye. This is very much in line with the theory of the Adjacent Possible. According to this idea of Stuart Kauffman, the Adjacent Possible is the set of “all” those things that are one step away from what actually exists and, hence, can arise from incremental modifications and recombination of existing material. The COVID outbreak has revealed a region of our collective Adjacent Possible where many options exist to combine smart and innovative working schemes with better mobility patterns.
Towards a Precision Mobility
It is with this spirit that, since last March, CSL Paris joined the Task Force of the Rome Agency for Mobility devoted to the COVID-19 emergency. The Task Force aimed at forecasting the mobility demand during the lockdown and its progressive release, with the objective to better tailor the public service and prevent disruptions and unsafe conditions.
The need for physical distance between individuals is a key variable today, encompassing the concepts of “safety” and “security”. The term security has also acquired the meaning of “possibility of minimising the risk of physical contact with potentially infected individuals”. Rethinking mobility in the “Postcovidic” era also means considering public health security.
Safety Distancing Index
The first contribution that we gave to this topic was the introduction of a new index, the Safety Distancing Index (SDI), meant to quantify the risk level while travelling on public transports. Public transportation is presumably one of the urban system’s components most affected by the safety distancing constraints. The need to reorganise transport services to face the current crisis is a challenging task that institutions are currently tackling. In this process, the emergence of potentially massive shifts in individual habits might be in contrast with the long-standing aims of reaching sustainable and equitable cities.
To define the SDI, we need to shift the attention from average fluxes of people, or average occupation numbers, to actual occupation numbers. If the maximal capacity of a bus is 100 passengers and in one hour the bus runs 20 times carrying 500 passengers, we can easily conclude that there are, on average, 25 passengers per bus. Suppose that 25 passengers can be a safety occupation within a bus. We know very well that fluctuations are always there and we can carry the same number of passengers in a myriad of different configurations on the 20 buses. In this specific case, there is only one safe configuration, i.e., the one where all buses carry 25 or fewer passengers. In all the other cases (and there are many of them, roughly 10^35 combinations), there will always be at least one bus with more than 25 passengers, leading to a potentially unsafe condition. In other words, to correctly compute the actual occupation number, we need to know who is on a specific transport means at any given time.
Fig. 1 – Schematic representation of the Safety Distancing Index (SDI). According to the travelling paths of individuals using public transport, some buses might be more or less crowded, even on the same bus line. An adequate physical distancing can occur only when the number of passengers is below a given safety threshold.
The SDI for a specific means of transport is defined as the fraction of travelling time an individual on that means of transport spends in unsafe conditions. The definition of “unsafe” condition depends, in turn, on the transport means capacity. In our case-study, a bus was considered unsafe if the number of passengers is higher than 25, a metro train unsafe with a number of passengers larger than 150 and a tram with a number of passengers larger than 50. A number of passengers higher than the values above for the corresponding type of means of transport will likely result in the impossibility to ensure physical distancing, possibly leading to unsafe conditions for the passengers.
The term Precision Mobility comes from the above constraints: knowing average fluxes, it is not enough to guarantee safety while travelling, and one has to move to a new level of sophistication. The computation of the SDI requires us to know how many people are simultaneously present on a given bus, metro, train, etc. This is different from the traditional approach of fluxes and Origin-Destination matrices. The method is called Lagrangian, and it consists of following the individual trajectories of people to reconstruct the complex space-time network of co-presences, i.e., being at the same time on the same means of transport. Notice that, in our approach, we never use actual individual trajectories. We rather reconstruct artificial trajectories featuring the same statistical properties of the real trajectories. In this way, our approach is fully compliant with the requirements concerning individuals’ right to privacy.
More in detail, the computation of the SDI requires both accurate and finely tuned models and the merging of information coming from a lot of different data sources. To this end, we exploited several data sources: (i) High-Precision Anonymised Location-Based (HFLB) de-identified data collected via mobile phones app (provided to CSL Paris by CUEBIQ in the framework of their Data4Good program); (ii) Census data about population and economic activities in different parts of the city; (iii) Public transportation data in the General Transit Feed Format (GTFS); (iv) boarding counts on public transport (in our case at metro stations). The mesh-up of all these data sources provided us with an unprecedented array of data that allows for precise quantification of the number of passengers travelling at the same time on the same means of transport (bus, metro line, tram, train).
Fig. 2 – SDI for the metro lines in the Rome Metropolitan Area as predicted by our scenario for May 18th 2020.
Visit our GitHub page for more details on the SDI in different hours of the day.
The notion of SDI is a key ingredient potentially affecting all future plans and analysis of mobility. For instance, one could think of a new generation of “paths navigators” optimised for safety and not only for speed. And, of course, SDI is highly relevant in assessing the goodness of “what-if” scenarios.
Forecasting scenarios
The second objective of our collaboration with the municipality of Rome was the development of a forecasting framework to predict the demand for mobility in specific days marking partial releasing of the lockdown measures. We worked on forecasting the mobility demand and the level of risk in two days, May the 4th and May the 18th. Our forecasting scenarios are based on the reconstruction of a database of about six million daily trips within the metropolitan area of Rome in the pre-COVID era (an area of 5352 squared kilometres with a population of around four million people): who is moving, where and when they departed from, where they are going, and why (work, education, leisure, returning home, etc.).
Starting from the pre-COVID situation, and based on the official restriction measures in place as a function of time, we constructed future scenarios by singling out the different components of mobility and their relative suppression (for instance, the reduction of workers moving due to the adoption of smart-working schemes, or the reduction of the education-related movements due to the closure of schools). For each scenario, we forecasted a typical day by inferring the occupation levels of each means of transport (bus, metro trams) at every time of the day and computed the corresponding SDI values. The ensemble of these predictions was released to the mobility agency of Rome and officially published before the date they were forecasting. In this way, we issued the forecasts without any bias connected to the knowledge of the real outcome. It turned out that our predictions displayed an overall quite good level of accuracy with extremely good predictions for metro lines and good predictions for the surface transports, like buses and trams.
Conclusions
The COVID-19 outbreak highlighted the need for scientifically validated scenarios that can support decision-makers. The point is not to find the “excellent” solution, but rather to provide the community with tools to conceive possible solutions and evaluate their effectiveness and opportunities. The space of possible solutions for mobility ecosystems is inherently complex due to the strong interrelationships between the different components of the system. One important source of complexity is coming from the demand for mobility and the whole universe that characterises it: who, why, where and when. Each of these “W’s” carries with it a micro-universe. “Who” implies the characteristics of the person who moves, her specific needs, his weaknesses. “Why” includes all the motivations for which one moves and brings with it the entire spectrum of work and relationship systems. “Where” and “when” refer to all the space-time constraints related to mobility. An additional level of complexity is provided by the offer, i.e., the set of means of transport that can support the demand for mobility. The interplay of demand and offer makes the space of possible solutions a complex landscape that has to be explored using reliable compasses. This is fundamental if we want to achieve an inclusive mobility environment able to improves accessibility and social inclusion, travel experience and citizen well-being.
Visit our GitHub page for more details on SDI and other articles on this topic.