Introduction

Rapid urban growth and related pressures on the global environment are challenging the governance and planning of cities1,2,3. Recent frameworks suggest various levers to bring about urban transformation towards sustainability4,5,6,7,8,9. However, urban planners and decision-makers struggle to implement the transformation processes in complex, real-world settings9,10. Effectively directing urban development towards more inclusive, resilient and sustainable urban systems11,12 requires multi-dimensional and radical changes13,14. Latest debates have pointed to the oversight of the ‘inner world’ of sustainability in these systemic views of transformation, including the emotions, thoughts, identities and beliefs of individuals driving human behaviour, otherwise referred to as a ‘deep leverage point’15,16,17. At the same time, there is a proliferation of data generated by massive ubiquitous sensing systems18 allowing to capture and monitor human presence, action and even intention19, yet imprinting themselves on our behaviours in unconscious ways, often steering unintentionally urban transformation20. This calls for a better understanding on how to harness these sensing systems for supporting the needed transformative change within urban planning.

Recent years have seen a massive development of sensors, increase of Earth Observation (EO) and geospatial mobile big data (such as mobile phone data or social media data), Volunteer Geographic Information (VGI) platforms, image analysis and motion detection. One-hundred and fifty billion networked measuring sensors are expected to be installed in 10 years and the Internet of Things (IoT) is forecasted to escalate data-driven decision-making19. These sensing systems promise many new opportunities to integrate knowledge about ecological, social and technological contexts of actions with the ‘inner worlds’ for triggering transformation21. For example, geographic information observatories now provide the potential to combine data about human preferences and behaviour data with biophysical data streams such as traffic counters, public transit, weather stations, news portals and air quality monitors22. Furthermore, smart building technologies increasingly integrate sensing systems in everyday objects23, and digital twins of cities are linked to multiple real-time data sources to allow citizen feedbacks24. Such integrated sensing systems can help generate a holistic perspective of places, regions or the entire globe, facilitating both observation, experimentation and prediction about people, processes and structures forming the city, and their changes.

Here, we investigate the different forms of sensing systems and their role in urban sustainability transformation. We conceptualize the interactions between various types of sensing systems and highlight their risks and benefits for stakeholders through four global urban case studies. We assess how deliberate sustainability transformations arise based on the interacting spheres of transformation suggested by O’Brien25, and identify which combinations of sensing systems support specific urban sustainability challenges. Based on these insights, we present a set of key principles to guide urban transformation processes towards ‘good’ Anthropocenes14,26.

Conceptualizing passive and active sensing

Passive and active sensing can be distinguished based on the types of technologies and level of stakeholder engagement. Passive sensing is generally associated with fast wireless communication, cyber infrastructure and the IoT, and the collection of real-time information without any active forms of stakeholder engagement27. This includes the collection of geospatial data about a given phenomenon using passive and active sensors (e.g. EO technologies, geospatial social media, mobile phone records), or the spatially explicit statistical data (e.g. population density measures). In contrast, active sensing technologies draw on voluntary contributions of people to collect geospatial data (e.g. survey research, Public Participation Geographic Information System (PPGIS), serious games). Active sensing aims to support consultation, engagement and empowerment of diverse stakeholders in urban planning through inclusion of the individual as both ‘being the sensor’28 and being sensed. It draws on the ‘wisdom of the crowds’ and public judgement in ways that provide spatially explicit information that can guide urban planning29.

To identify various entry points for informing urban sustainability transformation, we link the passive and active sensing systems to the three spheres of transformation: the practical; the political; and the personal sphere25. These spheres provide priority points for intervention to achieve transformation16,30. The practical sphere includes planning interventions that directly contribute to sustainability goals in cities25, and is mostly driven by entities directly operating or managing specific resources, such as planners, businesses or facility managers. The political sphere10 includes instruments (e.g. rules, incentives), institutions and community engagement processes for governing urban solutions, and is carried out by entities acting on behalf of wider societal groups or organizations, such as national governments, municipalities, NGOs and academia. In order to tackle complex urban sustainability challenges in city planning and corporate activities, we, however, need open and transparent data flows and approaches for assessing how the practical and political spheres combine with the personal sphere to influence transformation15. This personal sphere, driven by individuals, guides perceptions of transformation and choices about how we live in cities. Figure 1 shows the three spheres embedded within one another in a defined hierarchy: the practical sphere is at the core of transformation and is related to observable and measurable outcomes. Such responses are, however, highly dependent on the political, economic, legal, social and cultural structures associated with the political sphere. This sphere defines how and in what ways transformation at the practical levels could happen. The outer sphere with individual and collective beliefs, values and worldviews, frames the issues and the solutions that are addressed, and can be highly powerful as it influences the other two spheres. In the following, we provide examples for the various types of sensing systems (see Table in the Supplementary information for additional information on the risks and benefits related to these sensing systems).

Fig. 1: Combinations of active and passive sensing to inform urban sustainability transformation.
figure 1

A The active and passive sensing in isolation; B how active and passive sensing can be combined across the spheres of transformation (practical, political, personal) through the actions of eliciting, diagnosing, explaining, and predicting.

Full size image

Passive sensing of the practical sphere refers to the use of geospatial data informing decision-makers about key indicators of urban sustainability31 (e.g. land use, population density, building stocks32, service networks33 and energy potential34), and assessing changes in natural and human-induced processes (e.g. biogeochemical cycles, land cover changes, and climatic variability and change)35,36,37. Passive sensing of the practical sphere also includes discerning the patterns of human activities, such as changes in travel in response to COVID-19 outbreak38, detecting illegal fishing39 or monitoring forest fires40. Active sensing of the practical sphere in the context of urban transformation commonly focuses on the deliberate collection of geospatial data using, for example, VGI but can also include active urban experiments through Urban Living Labs (ULL) of specific places, through multi-stakeholder participation and user involvement, co-learning and co-evaluation and refinement41,42. Active sensing of the practical sphere can employ a suite of methods including interviews, surveys, envisioning workshops43, participatory modelling44, serious games45 and 3D simulation of landscape development46.

Passive sensing of the political sphere involves the use of social media and IoT technologies by politicians or value articulating institutions to influence civic opinion or guide certain behaviours47. For example, social media companies often seek to persuade and influence individuals and groups to take certain actions48 or to understand their civic and political participatory behaviour, both online and offline49. Others have started using IoT technologies to understand the complex evolution of legal systems8. Active sensing of the political sphere includes, for example, games that enable stakeholders and institutional decision-makers to assess the trade-offs associated with different sustainability and resilience policies50. They increasingly feature in climate change communication, participatory research and collaborative learning45 and can often be part of ULL activities.

Passive sensing of the personal sphere draws on a range of methods to collate data on the individual, including their values, attitudes, beliefs, preferences and behavioural patterns. Like passive sensing of the practical sphere, this sensing system can also draw on social media, telecommunications, EO technologies and other sources of big data, for example to assess changes in behaviours through the collection of geospatial social media data (e.g. through Flickr, Twitter)21,51,52. In contrast, active sensing of the personal sphere involves eliciting citizens’ behaviour and perceptions based on geospatial data collected and produced by volunteered citizens, planners and researchers in everyday living environments. It involves inviting citizens to express their values, preferences or behavioural patterns individually or in groups53using a variety of methods, e.g. interviews54, geographical ecological momentary assessments55, mail-based or online surveys, workshops, serious gaming56, participatory mapping57 and cognitive psychological methods58,59.

Combining active and passive sensing to inform urban transformation

We acknowledge that sensing technologies alone cannot trigger change. However, combined active and passive sensing (Fig. 1B) provides important benefits for informing urban sustainability transformation, compared with traditional urban planning systems, which treat these systems in isolation (Fig. 1A). We consider the transformational role of new combinations of active and passive sensing through a workflow including four separate actions of eliciting, diagnosing, explaining and predicting (Fig. 1B)60,61, which are anchored in a geographic context.

Eliciting involves identifying the different elements of each sphere of transformation relevant to the change process in the specific urban contexts. It is mostly conducted using passive sensing but is increasingly done in a hybrid manner using the active participation of local volunteers.

The diagnosing action involves exploring the interlinkages between the elements at a given point in time and space. It is also mostly driven by passive sensing but can be supported by the active participation of stakeholders to identify key entry points into the system. From an urban planning perspective, eliciting and diagnosing enable a more detailed understanding of the boundary conditions that influence transformation at a specific moment. Practical questions that can be addressed during this action include for example: ‘To which extent changes in behaviour have occurred?’, and ‘How do these passively observed behaviours match with stated preferences?’, or ‘What levels of agreement or disagreement about priorities for transformation exist across systems?’

The explanation action reflects on the sensed data by assessing the validity and uncertainty embedded in the intertwined results as a basis for managing risks associated with system transformation. Here, planners can obtain more detailed insights into issues of spatial data quality across system dynamics62, for example, differences in spatial accuracy between active and passive sensing data and different types of uncertainty linked to the integration of passive and active sensing data63,64,65. Planners can draw on both active and passive sensing systems to ask questions like ‘How well does the active and passive sensing data explain known changes in the urban system?’ and ‘What level of confidence can we assign to the combined results?’.

The prediction action aims to assess the potential for future change by integrating passive and active sensing into a modelling environment. It is the basis to define and design desired pathways of transformation66 and ultimately to inform institutional decision-makers about the needed governance adjustments. As long-term monitoring is usually missing in active sensing, predictions are usually based on models harnessing passive sensing. However, long-term monitoring involving the combination of active and passive sensing systems is crucially needed to better understand and predict drivers to and roadblocks of transformation within and across the practical, political and personal spheres10. Additionally, predictions that combine active and passive sensing data are needed to identify where and how new policy and legal options can nudge changes to existing institutional arrangements to dismantle the roadblocks and trade-offs to transformation.

The eliciting, diagnosing, explanation and prediction actions should not be considered in isolation in urban transformation processes (Fig. 1A). Rather, we argue that the results from each stage should feed into a dynamic relationship between the passive and active sensing (Fig. 1B), supporting knowledge co-creation processes and fostering dialogues and social learning about transformation opportunities and risks among scientists, planners, businesses, governments and citizens5,67. Such dialogues enable deeper understanding of how different modes of governance interact with transformations and the roles of routines, cultures, ideals and social groups in supporting or impeding transformation68.

We explore these differences with reference to four global cases, including Zürich (Switzerland), Singapore, Dar es Salaam (Tanzania) and Lahti (Finland). Short descriptions of the sustainability challenges experienced in these case studies as well as the mechanisms and processes harnessing sensing systems to support the urban transformation are provided below (see case studies in the Supplementary information for a detailed description). Table 1 provides then a frame of thinking about the risks and benefits of various types of passive and active sensing dynamics across the three spheres of transformation for various stakeholders.

Table 1 Typology of active and passive sensing combinations informing urban transformation, featuring benefits and risks of the combinations for stakeholders.
Full size table

The metropolitan area of Zürich is experiencing ecological, economic and social challenges in the fight against urban sprawl. A recent ‘control-and-command’ policy, directly addressing the extent, range and type of land use, fuelled ongoing social exclusion and gentrification processes by putting pressure on public services and spaces and accelerating the redevelopment of old housing stocks69. A bottom-up active sensing process engaging the local community with planners and authorities was launched to drive the urban transformation process towards inclusive and liveable neighbourhoods70. Passively sensed data and a postal survey helped elicit practical and personal factors hindering the transformation and provided the basis for a more in-depth diagnosis of the challenges related to the transformation but did not feed into the active sensing process of the participatory workshops. The uncoupled passive and active sensing processes led to a loss of key information needed for implementation, which slowed down the process, as several intensive pilot processes were necessary to gain the support of the citizens for implementing the spatial development plan.

As an island state, Singapore is highly depending on its natural ecosystems. While the economic benefits of urban development are regularly used in the top-down planning processes, comparable information for natural capital is missing and highly reliant on the values that local beneficiaries attribute to these assets. The Natural Capital project71 collected a wealth of passively sensed data and conducted several active sensing campaigns to elicit and diagnose the health of Singapore’s ecosystems and the potential supply of ecosystem services72. This information was integrated into a 3D virtual interactive platform73 to assist the Singaporean government agencies in decision-making and formulating key recommendations for the future management of natural capital. While there was an iterative process between the active and passive sensing, the sensing process ran in parallel to the traditional planning processes and only helped increase awareness for natural capital. Furthermore, the Natural Capital tool was fed by proprietary data belonging to the decision-makers, hindering its full use by planners.

In Dar es Salaam, rapid population growth is coupled with uncontrolled and geographically extensive urban development. At the same time, the city is vulnerable to several climate risks, particularly frequent flooding. Reliable and up-to-date digital spatial information of the city’s infrastructures, local assets and climate risks is a critical bottleneck74,75. Passive sensing through satellites and drones coupled with machine learning and artificial intelligence has been used to elicit and diagnose city’s infrastructures and physical environment. Active sensing in the form of participatory mapping by students and local communities complemented EO data into locally valuable data products with validated labels and attributes76. This has been a major change to the previously prevailing situation, where the data gap has been severe and planning and investment decisions had to be made without up-to-date information of the local situations. Furthermore, local ownership and capacities to react to flood risks on the basis of local knowledge have substantially improved, since the residents have been engaged voluntarily to map their own living environment75,76.

The city of Lahti aims to combine an ambitious sustainability strategy with high-quality participatory planning practises. To realize their visions, the city has developed an ongoing participatory master plan process77 that integrates land use and transportation planning within a large-scale participatory planning process. The city uses passive and the active sensing datasets that are stored in the same geospatial database, which is then used across all sectors in the city to provide data for planning processes. The active sensing data creates conditions for the city officials to make decisions and plans based on understanding the values, needs and behaviour of the citizens, and linking them closely with the characteristic of the physical environment and sociocultural context. The dataset enables planners from different sectors to diagnose and explain the challenges, and suggest planning solutions based on integrated passive and active sensing. This strong participatory approach has strengthened the social capital and empowered the local community to engage in the political process78 and has made it possible for planners to re-evaluate the city strategy goals where needed.

Based on the four global case studies, we argue that the use of sensing in informing transformation is being slowed down by a lack of coordination between active and passive sensing systems, hindering transformations across the three spheres. When passive and active sensing is uncoupled, the urban transformation process is often based on weak knowledge and overlooks available passive sensing, which can lead to manipulative actions by planners and regulating bodies in the active sensing process (Table 1, Case Zürich). However, even when there is active and passive sensing in place, they can be decoupled from decision-making, thus not supporting transformation despite the potential to do so (Table 1, Case Singapore). A dominance of passive sensing in a coupled passive–active system allows planners and businesses for effective urban transformation of the practical sphere, but requires building trust between data producers, data owners and data users (Table 1, Case Singapore). In the case of a dominant active sensing process, the passively sensed data is turned into locally valuable data products, increasing data ownership in individuals and local communities, and ultimately also triggering important changes in the personal sphere (Table 1, Case Dar es Salaam), but securing a long-term voluntary process requires engagement of governing bodies and a strong transformation in the political sphere. Finally, the Lahti case (Table 1, Case Lahti) shows that when active and passive sensing are used in a balanced manner, social capital can be strengthened and individuals and the local community empowered, fostering transformation in the political sphere and increasing stewardship of citizens for the process.

Reaching a full integration of the passive and active sensing in a dynamic process between eliciting, diagnosing, explaining and predicting, requires, however, not only an active exchange between the sensing systems, but also the integration of the practical, political and personal spheres (Fig. 2). Actors, institutions and processes need to be established in such a way to allow weaving multiple forms and systems of knowledge across planning sectors and transformation processes. Such a full integration would allow for different spatio-temporal dynamics and real-time assessments of the complex relationships and dependencies between the three spheres of transformation. This would address mismatches between fine and global-scale data through processes of internal and external validation by scientists, planners and citizens.

Fig. 2: Combined active and passive sensing loops across the spheres of transformation.
figure 2

The sensing systems are integrated in an iterative dialogue along the sensing loop.

Full size image

Principles for informing urban transformation using sensing systems

Building on the knowledge acquired in the case studies, we present a set of four principles guiding urban sustainability transformation and present ways forward for stakeholders to successfully realize transformations when harnessing passive and active sensing. We do not claim to define a theory of change per se but highlight the role of sensing as a basis for catalysing transformation in the three spheres.

Full integration of the spheres of transformation in an iterative dialogue along the sensing loop

When there is a two-way dialogue between passive and active sensing, goals are negotiated and knowledge is co-produced between the sensing systems, supporting a value-driven informed navigation towards sustainability (Case Lahti). Feedbacks from the active to the passive sensing help raise awareness of the triggered changes and ultimately improve the fit between the ecological, social and technical contexts, allowing adaptation in face of changes10. Focusing on ‘doing’ rather than just ‘reaching understanding’ is known to help integrate different perspectives and drive transformation79. In particular, design approaches, informed by passive sensing, can bring knowledge into a societal discussion and avoid the obstruction of emerging, more innovative and sustainable transformation pathways (Case Zurich). Co-design and co-production can however end in lock-ins when current pathways are not transcended, especially when governance systems, mindsets and urban infrastructures are rigid and create path dependencies (Case Singapore).

Commitment to new modes of governance enabling direct feeding of sensed information from the practical and personal sphere into the political sphere

Our case studies show that urban sustainability transformation can be fostered by activating individuals’ values, preferences and behavioural patterns; however, collective actions are also necessary to influence planning outcomes. New modes of governance are needed to dynamically integrate data flows from active and passive sensing systems in order to facilitate the formation of shared values and agreement on preferred actions regarding transformation towards urban sustainability80 (e.g. Case Lahti). Running the active and passive sensing processes in conjunction with an official planning process to allow feedback between sensing and decision-making would increase the relevance of the sensing process and help define necessary adaptations of policy and legal–institutional aspects to support sustainability transformation. Furthermore, voluntary engagement of local communities combined with automated mapping methods could offer cost-efficient ways to harness the sensing systems (e.g. Case Dar es Salaam).

Account for data biases and issues of representation and exclusion in the sensing processes

More data does not guarantee better evidence. Active sensing can help in giving value to passively sensed data and better define ownership for locals (Case Dar es Salaam), but differences in scale effects and data quality across sensing systems need to be carefully managed62. Further, attention should be directed to issues of ‘whose voices’ are considered or ignored in the planning process when drawing on passive and active sensing. Social and spatial injustice can be created by inequitable distribution of sensors across the area at stake81, reduced modularity and reduced interconnectedness through highly connected systems82. Striving for consensus regarding the selection and use of the data can lead to power plays, and requires active engagement to tackle the politics of sensing processes to avoid replicating existing power asymmetries83.

Promote high ethical standards, responsible research and openness of sensed data for the benefit of stakeholders

A major challenge is that amidst the open data movement, companies, businesses and also government agencies (Case Singapore) are increasingly becoming data owners (e.g. the mobile phone companies’ data on people’s mobility during the COVID-19 pandemic). This raises questions about how active and passive sensing systems can be open to the world and used for social good84. By ensuring access to sensed data, interested stakeholders can engage in evidence-based decision-making and the monitoring of development progress. An active participation of the data users to determine semantics and ontology of the data85, be it during an urban transformation process itself or in the Semantic Web community, is essential to communicate uncertainties and develop adaptive solutions. Data users need to understand how the raw data has been processed and for which purpose it has been collected (Case Lahti). In particular, data collected by active sensing of the personal sphere need to be coupled with informed consent and protected or anonymized to ensure privacy. Finally, the collection and handling of data needs to comply with the data protection rules relevant to each country and region.

Table 2 outlines concrete suggestions about how stakeholders can integrate passive and active sensing within and across the practical, political and personal spheres. Such mechanisms, integrated in a dynamic process with short cycles between passive and active sensing, will allow more rapid shifts in the practical, personal and political spheres, thus increasing the feasibility of actions within a given context10, and ultimately accelerating urban transformation. However, it is still unclear how such loops would fundamentally alter decision support systems and knowledge management. In particular, with the proliferation of real-time data-driven decision-making, there is an increasing tension between the practical and personal sphere. Research in cognitive and behavioural sciences will be essential to investigate how such integrated mechanisms modify individual’s perception for problems, use of information and the development of solutions, ideas and knowledge. Ultimately, it might require revisiting existing decision-making models. Furthermore, the privacy and security paradoxes, spanning the tension between a desire to collect individual data and the pursuit of anonymity, will require more knowledge in jurisprudence. Changes in data collection methods will have implications on equality of representation of different societal actors and data credibility, and might require new ethical codes for the use of the value-driven data. New data technology architectures integrating passive and active sensing will be needed and fitted to various types of governance. Innovative data ontologies and grammars will have to be introduced. Finally, new skills of decision-makers will be required to manage data and processes leveraging integrated passive and active sensing loops in dynamic real-world contexts.

Table 2 Mechanisms to integrate passive and active sensing within and across the practical, political and personal spheres.
Full size table

Conclusions

In conclusion, while the proliferation of sensed data promises unimagined new opportunities to transform cities towards sustainability, it will require a careful integration of the passive and active sensing systems. Integrating important insights from the political and personal spheres into the development of practical solutions will enlarge the solution space for responding to the global challenges. However, such system integration will need to be complemented by open accessibility to data and an active participation of data users in the choice of data products and related sensors. Information flow from active and passive sensing will need to be carefully coupled with existing urban planning and other decision-making processes in order to foster new understandings of choices, which can be made by citizens, planners and corporate stakeholders to facilitate transformations toward urban sustainability.