Data-driven policy has the potential to maximise positive outcomes on the environment, public health and the economy. In this blog, Prof David Topping, Prof James Evans and Dr Thomas Bannan explore the benefits of using big data tools as well as the challenges facing the implementation of data techniques to inform air quality policy in the UK.
- Real-time data from modelling tools such as ‘Digital Twins’ can predict the quantified impact of proposed air quality policy interventions.
- A shortage in digital skills in the public sector is the key barrier limiting the use of big data models to shape air quality policy.
- Local authorities and DEFRA should address this skills gap by training new employees and upskilling existing ones to improve digital literacy in the workforce.
The increasing amount of data collected on the environment and how we conduct our lives has the potential to dramatically transform policy design and implementation, particularly in the field of air quality. The size of the global datasphere is predicted to double less than every two years. Each day, thousands of monitoring stations around the world gather vast quantities of data on air quality of varying veracity and detail. Over a long period of time, the scientific community has developed modelling platforms that simulate the emission, transport and evolution of pollutants in the atmosphere. New, data-driven platforms – utilising near real-time data – offer a tremendous opportunity to predict quantified impacts of proposed policy interventions and thus determine interventions with the greatest outcomes on public health, the environment and the economy.
Big data in action
‘Digital Twins’ show great promise as a tool for informing air quality policy. A Digital Twin is a virtual counterpart of a given system. It uses near to real-time data to provide a snapshot of how the system is responding to multiple stressors. This can be used to assess how the system may respond to changes in conditions. What happens if a planned traffic intervention leads to increase traffic through another neighbourhood? Can we integrate transport data and personal mobility data into a near real-time model of personal exposure for population exposure estimates? These are the sorts of questions that can be answered utilising Digital Twins but might have otherwise been difficult to answer using traditional numerical models.
The smart city movement also provides numerous illustrations of the potential of big data-driven policy. Smart cities have been in vogue for many years, facilitated by the boom in the Internet-of-Things market (IoT). Managed static networks of sensors or distributed collections of personal devices monitor the environment in which people spend their time. Data from these networks is then used to design data-driven services for the public. For example, the monitoring of a transport system can include the use of image recognition cameras to detect vehicle type and volume. Ancillary data on public transport capacity, Bluetooth mobility tracking devices, parking spaces and public footfall provide a comprehensive picture on where and how the network responds to various stressors. This can be used to optimise traffic flows and minimise traffic-related air pollution around schools or hospitals.
There is also the potential to use the growing information about our air quality to better inform people about potential impacts of the choices they make day to day. There are already a multitude of apps that can give expected levels of air quality that an individual might be exposed to, which may influence how that individual travels to work, perhaps spending more time walking or cycling on clearer days. Likewise it might influence the route taken on a commute or school run.
A critical shortage in digital skills
Air quality is complex, and results from a variety of interdependent factors. A significant evidence base is required to build accurate models. For this evidence base to be useful, we need a skilled workforce across national and local governments that is capable of designing, implementing and analysing big data models.
The key barrier to UK implementation of big data techniques for air quality policy is a digital skills shortage across the public sector. To take advantage of the opportunity afforded by modelling platforms such as Digital Twins, we need trained specialists with the capacity to help drive digital innovation. In general, local authorities do not currently employ enough of these professionals, and the analysts they do employ often almost exclusively focus on statutory reporting.
Bridging this skills gap should be a specific priority for Defra as well as local authorities responsible for air quality. Data innovation specialist roles must be established. To fill these roles, there must be an increase in targeted recruitment from the existing pool of skilled workers in the UK. This pool of data specialists must be expanded, both through upskilling current employees and through equipping the UK population with improved data skills.
The recent Skills for Jobs white paper emphasises importance of digital skills in the modern workforce; progression to advanced technical study is a core purpose of the proposed national skills and educational reforms.
In the National Data Strategy, the government commits to training 500 analysts across the public sector in data science by 2021, through the Data Science Campus. Manchester’s Digital Strategy is a fantastic example of a local authority digital framework that puts innovation and ‘smart people’ – an upskilled, digital literate population and workforce – at the forefront. There is already a great deal of misinformation and less than trustworthy data available, especially regarding air quality sensors, and a more digitally literate procurement strategy at both local and regional levels would help by not only saving money, but also avoiding decisions being taken based on less rigorous data.
The Fast Track Digital Workforce Fund in Greater Manchester is a great scheme that seeks to address locally identified digital skills gaps, and provides accessible pathways into digital careers for underrepresented groups. Such programmes should be supported and expanded to offer specific training on Digital Twins, with a view to expanding to a national Data Twins training programme. Specifically, it is vital that as developments of hybrid and machine learning driven representations of our air come from various sectors, there should also be training in validation protocols and a fundamental understanding of model construction. Local authorities are already operating under immense budget pressures, and connections and partnerships with Universities are a clear opportunity for a mutually beneficial relationship.
Other barriers to big data
Critical digital infrastructure is required to support Digital Twins, with open standards playing a central role. Simply put, we need to gather a lot of data and the data needs to be of a reasonable quality. Unfortunately, there is a considerable disconnect between commercial low-cost air quality monitoring solutions designed for dense networks and the technologies used in an academic setting. Research grade instrumentation not only captures information on pollutant concentrations and trends but quantifies a chemical signature of our atmosphere. This information is important in resolving changes in emissions and processes. However, the gulf between the financial entry point and ease of erecting dense networks of commercial solutions rather than academic tends to dictate scalable network design. This includes not only the cost per unit, but required installations, maintenance, quality control checks and decommissioning.
There are significant concerns surrounding quality control, ownership and differences in calibration procedure for commercial air quality monitoring equipment. We must find solutions that combine the academic rigour of facilities traditionally maintained by the academic community with low-cost network solutions. A framework must be developed that mandates benchmark equipment quality, outlines a common calibration procedure and ensures transparency over equipment and data ownership.
Mapping air pollutants and modelling the chemical and physical processes they undergo allows us to predict quantified impacts of proposed policy interventions. Data-driven modelling and Digital Twins may well be the most efficient route to decision making in an evolving environment. To deliver on this tremendous opportunity, we need to grow the evidence base surrounding air pollution and set firm standards for its quality; we must also upskill the workforce across local authorities and Defra to bridge the digital skills gap.
This article was originally published in On Air Quality, a collection of thought leadership pieces and expert analysis on how to tackle air pollution, published by Policy@Manchester.
Policy@Manchester aims to impact lives globally, nationally and locally through influencing and challenging policymakers with robust research-informed evidence and ideas. Visit our website to find out more, and sign up to our newsletter to keep up to date with our latest news.