How much pollution are we exposed to in the city? Measuring pollution exposure using mobile network data
The number of people residing in cities continues to grow at an unprecedented rate. This rapid urbanisation is causing pollution exposure levels to exceed recommended limits. How can we address this challenge? The first step is to truly understand its impact. Dynamic data sources, such as mobile network data, enable us to accurately measure how this pollution affects the population, empowering us to design effective public policies for healthier, more sustainable cities.
According to the UN, over 50% of the global population currently lives in cities, and it is estimated that by 2050, this figure will exceed 66%. This growing urbanisation leads to a higher concentration of polluting emissions, exacerbating pollution levels in urban centres.
This phenomenon is largely due to increased traffic, but also to the growing energy demand in the residential, industrial, and service sectors, which significantly increases their emissions.
The consequences of this pollution directly affect the health of city dwellers, particularly in the long term. Specifically, according to the World Health Organization (WHO), air pollution is the second leading risk factor for non-communicable diseases (behind only tobacco), causing 4.2 million premature deaths worldwide in 2019.
Urban pollution: a public health challenge
Given this situation, cities have a responsibility to implement measures to mitigate these impacts. Key strategies include developing sustainable urban planning policies to incorporate more green spaces, promoting renewable energies, improving public transport, and adopting initiatives to reduce emissions from the most polluting sectors (such as implementing low emission zones).
However, to design effective policies that combat these effects, it is essential to have precise information on the distribution and impact of pollution. In this regard, two types of data are essential: the concentration of pollutants in the environment and population distribution. Understanding the population’s level of exposure to pollutants and how these affect them will enable more effective and efficient decision-making to improve air quality in urban environments.
Limitations of static models: Why do we need dynamic data?
Chemical transport models allow us to estimate pollutant concentrations in a region. Based on emissions data and meteorological conditions, they simulate the dispersion of harmful substances, making it possible to identify areas where they will concentrate. These models are key to estimating the presence of the main pollutants related to respiratory health, such as ozone (O3), nitrogen dioxide (NO2) and particulate matter (PM10 and PM2.5). Their application provides us with crucial information, as the location where a pollutant is emitted does not necessarily coincide with the location where its concentration will be highest.
However, understanding the true impact of urban pollutants requires more than just measuring regional exposure: it requires understanding dynamic population mobility. Historically, studies have relied on a static approach, analysing pollution levels solely based on residential addresses. This assumes citizens remain at home around the clock, a critical oversimplification that ignores the reality of daily urban movement.
In order to understand the impact of these pollutants it is necessary to know not only a region’s level of pollution exposure but also its population dynamics. Until now, many studies have adopted a static approach, analysing pollutants solely at people’s places of residence. However, this approach assumes that the population remains at home for most of the time, which is a critical oversimplification.
In order to conduct representative and reliable pollution exposure studies, it is necessary to understand how the population moves to find out where they are most exposed.
In this context, dynamic data sources, such as mobile network data, take on fundamental relevance. This data provides longitudinal information on the population’s mobility patterns, allowing for a detailed analysis of people’s locations and movements throughout the day. Furthermore, it includes users’ socio-demographic information (such as age and gender). Meanwhile, its longitudinal analysis makes it possible to identify relevant user information, such as their place of residence and work, or purchasing power. From this information, the purpose of a trip or activity (e.g., commuting to work) can be determined, which makes it possible to characterise the mobility patterns of different population groups.
By cross-referencing mobility pattern information extracted from mobile network data with pollutant concentrations obtained through chemical transport models, it is possible to more accurately analyse the population’s actual exposure throughout the day. This allows us to know not just the pollution levels at each location, but also how and when people are exposed to these pollutants.
This information is essential for evaluating the real impact of pollution in a region and determining who is most affected. Knowing the quantity of pollutants the population is exposed to, as well as where this exposure takes place, is key to designing effective policies that reduce its impact. Thanks to this data, it is possible to identify which areas require priority intervention to reduce the risk to citizens.
The advantages of mobile network data in air quality analysis
Using mobile network data as a dynamic source for studying pollution exposure offers several strategic advantages.
Bias correction
Firstly, characterising population dynamics makes it possible to correct the common biases of estimation following a static approach. For example, the exposure of people living in highly polluted areas is often overestimated, as many of them regularly travel to areas with lower pollutant concentrations. Conversely, the exposure of those living in less polluted areas tends to be underestimated, as their journeys often take them to areas with higher pollution.
Representativeness and socio-demographic precision
Secondly, the high penetration of the sample across all social strata ensures the representativeness of each population group when segmenting mobility patterns according to socio-demographic characteristics (age, gender, place of residence, purchasing power, etc.). This representativeness facilitates the identification of particularly vulnerable groups and the design of public policies to protect their health, aligned with environmental justice principles.
Segmentation by transport mode and activity
Furthermore, fusing and enriching this data with other sources, such as GPS traces or mobility surveys, makes it possible to differentiate between indoor and outdoor presence, as well as to segment mobility by transport mode. In this way, it is possible to more accurately estimate the exposure resulting from journeys made in different transport modes, identifying, for example, the exposure caused by mandatory mobility when commuting to work, as well as exposure at home.
Likewise, this data can be used to analyse the correlation with the incidence of cardiorespiratory diseases. An approach based on more detailed data allows for a better understanding of the real impact of pollution on public health and the design of more effective strategies to mitigate its effects.
Continuous updating for decision-making
Finally, using mobile network data also makes it possible to measure whether a specific population group’s level of pollution exposure is within the ranges established by the WHO. Its precision and versatility offer an up-to-date and realistic picture of the situation, facilitating the identification of issues that require urgent intervention.
Success stories: Nommon’s commitment to sustainable urban planning
Nommon is involved in several projects to analyse population exposure to pollution in different locations across Spain.
The SIMAD project, led by Nommon with the participation of Aimsun and the Universidad Politécnica de Madrid, has developed a tool for monitoring air quality in the city of Madrid. This tool monitors emissions, concentrations, and the population’s exposure in the region, as well as simulating future scenarios. This facilitates an integrated analysis of all three aspects with the aim of defining effective measures for developing more sustainable cities.
The MePreCiSa project, led by the Barcelona Supercomputing Center in collaboration with Nommon, sought to provide a flexible and scalable cloud-based solution for the integrated analysis and management of health issues in cities, through the integration of population mobility data based on mobile networks, health data, and environmental data. One of its use cases focused on measuring the impact of air quality on the population of Catalonia, with the aim of understanding the relationship between mobility, air quality, and public health. The results of this analysis will serve to support decision-making in the design and implementation of effective interventions that improve the health and well-being of the inhabitants of Catalonia.
Both projects leverage Nommon’s Mobility Insights and Population Insights solutions to analyse population dynamics within a territory. These solutions make it possible to extract the mobility and activity patterns, respectively, of a region’s population. Analysing this information alongside pollutant concentration data provides a detailed view of the population’s exposure to different pollutants (O3, NO2, PM10 y PM2.5). Based on this, we can characterise this pollution exposure according to socio-demographic variables such as age, gender, and place of residence. This approach provides key answers to questions such as what level of pollution residents of a given area are exposed to, where this exposure occurs, which areas present the highest levels of exposure, and which social groups are most vulnerable to pollution.
