City Science: A Chaotic Concept – And an Enduring Imperative

Amid widespread enthusiasm for ‘scientific’ approaches to reading and understanding the city, driven by the increasing availability of ‘big data’ and sophisticated tech-driven analytics, we call for greater clarity in the terms and genealogies of the ‘new’ sciences of the urban. We appeal to those observing and participating in the latest manifestations of ‘city science’ to recognize the intellectual and practical inheritance of their urban craft. At the same time, we call for planners and social scientists, who would be critical of the emergence of city sciences, to move beyond longstanding debates of previous decades, and to engage more meaningfully and productively with new analytical approaches to understanding and intervening in urban processes. It is only through sustained intellectual engagements that we can hope to develop the new sciences of the city as an art, and to respond effectively to the global nature of the urban challenges that confront us.

The imperative of providing a summative account of the global changes generated by and impacting on cities has catalysed debate on how best to understand urban systems. Multiple articulations and interpretations of the ‘global urban’ and calls for a ‘new science of cities’ echo the passage of transnational policy prescriptions that include but are not limited to an urban Sustainable Development Goal. The race to reconfigure planning heuristics to prioritize urban transformation in order to address multiple global crises is already understandably, if possibly counterproductively, contested.

Debates surrounding the ‘new’ city sciences are polarized. On the one hand, a new generation of tech-savvy data scientists, spatial modellers, and analysts confidently express their ability to predict and explain city processes at unprecedented scales of complexity. On the other hand, those trained to see the world as fundamentally shaped by contingent meanings and subjectivities may see in such approaches little more than old positivism in new bottles, or perhaps a hubristic overstep of urban non-specialists onto their turf (Derudder & Van Meeteren, 2019). In some ways, then, the discourses of ‘city science’ invert Ian Hacking’s (1999) reflection on the celebrity of the term ‘social construction’:

The phrase has become code. If you use it favourably, you deem yourself rather radical. If you trash the phrase, you declare that you are rational, reasonable, and respectable. (p. vii)

Might not a dismissal of ‘city science’ have something of a radical feel to it; a statement of commitment to a certain kind of criticality and (certainly anti-neoliberal!) politics? Or, when used favourably, does one not mark oneself as grounded with a bloodhound fidelity to data, or even as (that rare breed) a believer in human progress through technological mastery? However, perhaps now is not the time to be taking the moral or theoretical high ground. Given the imperative to (simultaneously and immediately) identify the global contribution of cities to a dramatic reduction in emissions, to better health and wellbeing, to improved ecosystem integrity, to reduced economic and social inequality, and to more vibrant democracies and less corruption, heuristics that advance change at scale cannot be dismissed out of hand, even while ongoing critical reflection is sustained.

It is our contention that a polarization in the terms by which city sciences and their potential are discussed is thoroughly unhelpful if we are to develop a global understanding of urban challenges and how to confront them. There may be ways we could avoid sinking into such an impasse in the treatment of urban sciences. The first is to be mindful not to conflate, stereotype, or make monolithic what is in fact a wide domain of urban scientific practices and their established, if disparate, contributions to planning. A second device to foster dialogue on the value of the city sciences is to recognize the historical inheritances through and upon which they have emerged and thrived – in the process, contributing a great deal to planners and planning as a field. Recognition of these ambiguities might encourage would-be critics to engage the urban sciences in a more productive and caring spirit.

‘City science’ is not coherent or new, but it has endured and revitalized itself over time

It may be helpful for critical urban scholars to interpret ‘city science’ as a ‘chaotic conception’. Here we recall Rose’s (1984) critique of ‘gentrification’ as a term assumed to refer to ‘a single or unitary phenomenon’, yet in reality encompassing a wide range of different kinds of realities, actors, and causal processes (see also Sayer, 2010). By invoking the term ‘chaotic’, we do not seek to deny the utility of a ‘city science’ but rather to draw attention to the fact that, in its current manifestation, it is neither new, coherent, nor deliverable. This is not a unique charge. The discipline of urban studies is host to a number of concepts (not least that of ‘the urban’ itself) that are used in referring to a range of phenomena and processes spread across very different contexts, often without clear application (Brenner & Schmid, 2014). Such include ‘neoliberalism’, ‘informality’, ‘suburbanism’, or ‘peripherality’ (Harris & Vorms, 2017). The argument can certainly be made that the capaciousness of these terms, although irritating in their diverse invocations, offer a way to draw many kinds of places and dynamics together within a single frame. Of course, there is a danger that by using such vocabularies we end up sketching ourselves ‘a consoling image of how the world works’, glossing over important issues and ‘pluralistic differences’ that would enrich our emerging understanding of the ‘global urban’ in all its pathways and instantiations (Barnett, 2005; Parnell & Robinson, 2012). Can using the term ‘city science’ to refer to a range of practices and orientations surrounding data collection, geospatial modelling, statistical analysis, and ‘smart cities’ not obfuscate the plurality of methods and techniques (not to mention ethics and politics) active in this area; their interactions, their contradictions, and their differing potentials for disruption? Might we not, spurred on by the call to pursue a city science, overlook the precise and variegated ways in which new ‘pictures of the city’ are descending and emerging (Batty, 2016)? We would argue so.

Despite the range of critiques that have been aimed at city or urban sciences – some of which are outlined below – we note that the drive to provide a ‘big picture’ or set of pictures of cities and city systems is an enduring agenda. Scaling up (temporally, spatially, politically, sectorally) and understanding the interacting dynamics and feedback loops within and across the urban system is an undeniable imperative for urban researchers of all critical, theoretical, and methodological orientations to engage and grapple with if we hope to deliver on the promise of sustainable urban-centred futures. In facing up to the urgent demand for new ideas on how to make cities work better, it may be that learning from the past can be as important as pursuing novel modes of enquiry.

The most recent iterations of city or urban science are associated with recent developments in how we understand, explain, and predict city processes using computational modelling and simulation; approaches linked to the increasing availability of big data and the refinement of techniques such as machine learning (Kitchin, 2016). However, we would do well to remember that there have been historical waves of interest in and enthusiasm for scientific and data-driven approaches to urban or territorial/spatial issues, and that these have taken different forms and had varied impacts. In the first half of the nineteenth century, for example, the ‘avalanche of printed numbers’ described by Ian Hacking (1982) emerged from and helped generate a craze for a numerical analysis of the dynamics of population and the social problems linked to rapid industrial urbanization. Meanwhile, Charles Piquet and John Snow’s pathbreaking geospatial analyses and visualizations of cholera outbreaks in Paris and London laid the foundations for the application of spatial analysis in urban epidemiology. Later, Charles Booth and Benjamin Rowntree’s research into urban poverty in Great Britain helped to foster a more statistical yet interdisciplinary impetus to the study of urban problems. In the twentieth century, Walter Christaller’s central place theory developed a geospatialized and scientific (albeit less quantitative) approach to the study and planning of settlements, helping to establish the notion that urban systems could be analysed in terms of repetitive modular patterns and geometric hierarchies – an insight later taken up by computer-based modellers (Batty, 2012).

The resurgence of positivism and the momentum of the quantitative revolution in spatial and geographical sciences in the 1960s also enabled a series of movements towards scientific modes of urban analysis. Criticisms of the ‘fixities’ of earlier physical master planning practices, for example, fostered an enthusiastic uptake and application of systems analysis to urban and regional planning (Burrows, 1980; Davoudi, 2012). Here cities were seen as machinic, organized from the top down; ‘distinct collections of interacting entities’ usually existing in ‘equilibrium’ with core functions that allowed for the control of urban growth and change (Batty, 2012, p. S9). Factorial ecologists, drawing upon earlier traditions of urban ecology and social area analysis, tried to simplify and examine the fundamental relationships between the main elements of urban socio-spatial structure using multivariate statistical techniques (Berry, 1971; Berry & Rees, 1969). Moreover, for adherents to ekistics theory, or the ‘science of human settlement’, as for systems theorists, regular patterns existed in the relationships between objects that could be ‘mapped, modelled, and used as the basis for predicting future patterns’ – the complexity of settlement might then be distilled into ‘orderly classifications of size, location, and function’ (Davoudi, 2012, p. 434).

Meanwhile, developments in computer technologies led to a growing application and transformation of the term ‘model’ within urban and planning studies. Where prior to the 1950s it generally referred to architectural representations of physical urban forms, the use of mathematical models came to dominate during the 1960s and 1970s. In the process, a ‘model’ increasingly signified a symbolic ‘simplification’ of complex urban systems rather than a scaled-down ‘picture’ of reality (Batty, 2001). Technological advances of the 1960s further enabled the emergence and application of geographic information systems (GIS) for spatial analysis and visualization, developments later critiqued for entrenching a positivist epistemology (Schuurman & Pratt, 2002).

The pushback to these later rounds of enthusiasm for more ‘scientific’ planning methods emerged from various sources and inspirations. Critics of systems theorists argued that cities could not be studied as discrete enclosed systems with a benign environment and a predisposition to equilibrium (Batty, 2012). Political economists in particular critiqued a ‘naïve positivist empiricism’ as conducive to the maintenance of an unjust status quo (Harvey, 2009); quantitative geographers were charged with ignoring the underlying social and economic processes through which groups are constituted and resources distributed among them (Healey, 1986). Hermeneutics offered a popular alternative to the ‘crisis of knowing’ associated with the unbundling of positivist certainties; its proponents emphasized the discursive and dialogical nature of knowledge, in the process reframing planning as more of a ‘craft’ than a ‘scientific endeavour’ (Friedmann, 1987). The interpretivist position of poststructuralists shifted attention to the discursive and linguistic production of contingent urban concepts and realities, as well as the inherent relationality of space and spatial properties (Davoudi, 2012). Feminist and later postcolonial writers pointed to the dominant masculine and Eurocentric tropes that privileged the bodies, perspectives, and systems of thought of some groups and cultures over others in the ways that scientists conceived of and examined the city (King, 2000; Sandercock & Forsyth, 1992). For planning theorists, meanwhile, a postpositivist agenda enjoined a rejection of the assumed dualism between facts and values, as well as the positivist ‘distinction between substance (analysis) and procedure (process)’ (Allmendinger, 2002, p. 84). However, by the late 1990s and 2000s, planners had started to talk with renewed enthusiasm for a science of cities – one transformed by new developments in complexity sciences, now intent on analysing and describing urban systems as emergent bottom-up processes (Batty, 2012).

A key point that we wish to highlight here is that with each reinvention of city science, an enduring tension existing at its core is never properly negated or resolved. That tension is the relationship between heavily quantitative analysis and a wider, more robust framing and understanding of cities and urban systems on which planning rests. Some scholars have tried to move past a bifurcated approach. Probably the most sustained effort has emerged from political economists who deftly combined empirical and theoretical analysis to explain the production of space and inequality under conditions of capitalism (Harvey, 1978; Smith, 1990). Critical realists have offered a different approach and solution, but with some exceptions their ideas and methodologies have not proven to be particularly influential in urban and planning studies (Banai, 1995; Crankshaw, 2014; Næss, 2015; Næss & Jensen, 2002).

Despite these efforts, the ways in which the aforementioned tension is approached is often schizophrenic. On the one hand, some accept that those working in the urban sciences should be ‘left to do their own thing’. On the other, ‘external’ critiques by those who may not have a stake in the future of the field might seek to demonstrate what is problematic or ‘wrong’ with data-driven approaches to urban scientific analysis (Schuurman & Pratt, 2002). They may distance themselves from what is perceived to be a ‘fetishization of data’ and ‘idolization of method’ (Mattern, 2013). These critiques can fall into an oppositional binary that risks reproducing tired and unproductive debates pitting positivists against interpretivists.

Sciences of the city are contested, and may be undeliverable – but they are critical

We have argued that a polarized and sceptical discourse surrounds the present iteration of city sciences. In one corner, some urban scientists evince an enviable self-confidence that new analytical techniques and big data will comprehensively capture the essence and future of the urban. In concert, new epistemological positions have arisen. An inductive empiricist version goes so far as to pronounce the death of the traditional hypothesis-testing scientific method, and of theory – understanding correlation, it is argued, now and increasingly supersedes the analysis of causation (Mazzocchi, 2015). Another position holds on to the scientific method but considers that hypotheses should emerge from the analysis of data rather than from theory (Kitchin, 2016). In the other corner, we find a degree of handwringing over the very idea of the city sciences, with their association with big data and smart urbanism, and their prospective dominance within urban analysis and planning. Some hold that data analytics should merely follow and test the hypotheses and insights born of detailed and critical social scientific study. Or, one might find a stubborn refusal even to entertain the ideas and practices of those perceived to be ‘naïve positivists’. The debate is a cul-de-sac.

Planners and urban specialists cannot afford to be drawn into and reproduce this binary. From diverse places and perspectives calls are now being made for the creation of something more inclusive, systematic, and coherent in the ways that we approach scientific urbanisms (Acuto et al., 2018; Kang et al., 2019; Keith et al., 2020; Kitchin, 2016; McPhearson et al., 2016). Whatever forms this could ultimately take, we need to proceed in a way that avoids reproducing tired and unproductive debates and dualisms. Organizationally, this is not simple. There is no professional body, no single discipline or IPCC equivalent, for cities. We have no conduit or platform for city sciences; no place ‘where it all comes together’. It is a fractured space, conceptually and institutionally. Nonetheless, there is a demand for unprecedented urban interventions to mitigate global change and ensure that cities can adapt to planetary emergencies.

Notwithstanding the possible limitations of the city sciences, there are very good reasons for us to talk across cities, to scale up and secure a scientific tracking of urban drivers and impacts related to climate, biodiversity, health, and so on. We need to do more than look at objects, processes, and problems in isolation. Yet our city sciences have to consist of far more than simply urban data analytics, or among other things, we will fail to capture the long-term temporality of urban change, to influence meaningfully the design and planning of cities, and to reach places and issues that are cast in the shadows of big and routinely generated databases (Batty, 2019; Kang et al., 2019). In articulating this broader remit, city scientists have to be more modest in acknowledging their contradictions, historical legacies, and political and ethical assumptions. They must recognize that city science is not a clearly bounded concept or field, that it should be an open rather than a closed community, and that it should be ecumenical rather than fundamental in its paradigmatic approaches.

Meanwhile, those who are not directly involved in the urban sciences should approach their critiques of the field through a recognition of its variegations and subtleties, and a far deeper understanding of its substance, systematics, and technicalities. Experts in urban governance, politics, culture, and ecology can add much to projects of urban analysis, modelling, and prediction, particularly when these tools are applied in contexts that differ from those in which they were originally developed and refined. But doing so requires working with, rather than judging, the value of more technical analyses.

However, we also need something more than invigorated cross-disciplinary conversations if we are to develop the science of cities as an ‘art’ (Mattern, 2013). We can recognize at least three broad imperatives. First, we require robust conceptualizations that accurately describe and make legible the emerging state of ‘the urban’ while directing our attention to the dilemmas of technological urbanisms. The notion of the ‘machinic city’ is one effort to think critically of cities as spaces of flows and circulations, silently sorted and shaped by technologies (Amin & Thrift, 2002). In recognizing the ways the ‘machinic intelligence’ of cities now defines and mitigates risks (Amin, 2013), we are further called to consider how the technical DNA of urban life is structured and how it engineers the social and political hierarchies of cities. Examining how social values are written unevenly into algorithms, surfacing and confronting the ‘abuses of software’, and opening up debate on alternative ways of ‘weaving technology into the urban social’ (Amin, 2006, p. 1014) generates intellectual opportunities for critique and reflection, not just the production of code.

Second, in procedural terms, we might find inspiration from citizen science movements in how we approach an urban scientific craft. Such movements consist of far more than hackathons and eager ecological observers armed with smartphones. They include the community-based census and mapping exercises of organizations like Slum/Shack Dwellers International; initiatives that call attention to important issues of data ownership and partnership, enabling citizens to make visible their lived realities and express their needs to authorities, while providing up-to-date data for planners (Hachmann et al., 2018). In recognizing a more distributed and organic domain of efforts to articulate different kinds of knowledge in and of the city, we may locate a necessary corrective to expert-driven discussions around urban and natural systems (Townsend, 2015). Committing to a wider democratization of urban knowledge production and agglomeration could, if designed effectively, provide a common ground for both sides of the city science debate – a meeting point for renewed urban academic and governance practice that is simultaneously political, technical, and creative.

Third, in critical terms, practising an art of city science calls for engagement in a spirit and form of productive critique that expresses care for the subject (Schuurman & Pratt, 2002), recognizing that the urban sciences do not exist on the far side of a fundamental rift. Critiquing and flogging a broad set of ideas and practices for their supposed faults means that we simply end up with the same ‘mediocre biases’ with which we began. Here Graham Harman’s reflection on critiques surrounding the human/world dualism in philosophy seems apposite:

If we make no concessions and play along with nothing, then our ‘critical’ claims merely endorse one of the prefabricated positions of the day, whichever one it may be. (Harman, 2009, p. 119)

There is both a political and scientific imperative for developing a more open and coherent field of city sciences. Indeed, the ‘science of cities’ is an enduring concept not simply for intellectual reasons. In fact, the intellectual challenge of finding effective ways of working together is probably harder, because those working from incommensurate intellectual disciplines and traditions do not and probably will not agree. The imperative is more about the state and future of the planet which, increasingly marked by conditions of urbanity, demands that we find a way to work together.

Acknowledgments

This piece was completed with support from the PEAK Urban programme, funded by UKRI’s Global Challenges Research Fund, Grant Ref: ES/P011055/1. We would like to thank Edgar Pieterse and Levi Wolf for comments and suggestions on an earlier draft of this piece.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

James Duminy

James Duminy is Lecturer in Human Geography at the University of Bristol and Honorary Research Associate at the African Centre for Cities, University of Cape Town. A planner by training, his work examines the governance of urban change in the global south.

Susan Parnell

Susan Parnell is Global Challenges Professor of Human Geography at the University of Bristol and Emeritus Professor at the University of Cape Town, where she was a cofounder of the African Centre for Cities. She has been actively involved in local, national, and global urban policy debates surrounding the 2030 Sustainable Development Goals and is an advocate for better science policy engagement on cities.