Smart inputs, smart outcomes

A Smart City design process is one that employs urban technologies to create effective outcomes for people. Much attention has been paid to technologies that collect, visualise and analyse data. Less discussion has taken place about the use of data to design cities and buildings using predictive models. The Space Syntax approach combines data sensing, mapping and analysis with creative design. Experience suggests that data-driven design processes will become more popular in the planning of future cities. By Tim Stonor 

The future of cities will be decided by answers to key urban policy questions: Where, and in what numbers, are people going to live, work and take leisure? What forms of movement will exist to connect people together? What flows of resources will need to occur - for example with energy, water and waste – to support human behaviour patterns? What impacts will these flows have on the natural environment? 

Addressing a legacy of failure

 

Rapid global urbanisation provides the opportunity to create striking, new forms of city living, supported by the proliferation of new 'smart' technologies. Nevertheless, the enthusiasm of architects, planners and transport professionals should be set against the sobering reality of recent urban history. From pedestrianised precincts to 'culs de sac' and upper level walkways, many innovations in town planning and design have been launched with great optimism over the years, only to create blight in the form of massive social and economic cost. 

Past results matter greatly for the future planning of cities because professional failure creates public concern and this affects political confidence. So, before beginning to create visions for novel urban futures, those responsible – the architects, planners and engineers as well as the policymakers and politicians – should first ask why past plans didn’t deliver what was intended. Why, in seeking social harmony, wealth and resilience have cities become divided, unequal and car-dependent? 

There seem to be two key reasons for urbanism’s inability to accurately forecast the effects of its actions: first, the scarcity of real knowledge about how people behave in cities and what the impacts of these behaviour are on, for example: land use resilience, land value and safety. Second, the shortage of accurate and reliable forecasting tools to test plans in advance. Until recently it has been expensive and time-consuming to overcome these issues: gathering data through manual, on-site surveys using teams of observers with pens and clipboards has been costly; transcribing video captured by cameras into maps and statistics has been a slow and intensive process; computational power has been inadequate and expertise limited. 

A new 'science of cities'

However, the rise of the 'smart' era has witnessed an explosion of data capture and analysis techniques, as well as the technology-literate operators trained to use them. The opportunity exists for public, private and community-based organisations - for the first time in human history - to behave differently: to use automatic data capture, visualisation and analysis techniques to build databases of urban performance, analyse the patterns within these, then create future plans in a more robust manner (Reference 1). 

The prospect of a new science for cities is real: a new, evidence-informed and analytic approach to urban planning and design that might obviate, or at least reduce, uncertainty in future urban decision-making. 

Such an opportunity raises fundamental questions about what datasets should be collected and how these should then be analysed. Fundamentally, the ways in which built environment data is used must change. Architects need to think at a broader scale than before, planners at a finer grain and both professions need tools that fit these purposes. Currently, architects use Building Information Modelling (BIM) systems that handle data at the building level. While BIM can stretch to small clusters of buildings, it does not usually allow buildings to be set in their wider urban contexts. As a result, the important influence of context on place is lost and too many buildings are designed in isolation, with obviously negative results once built: turning their backs to each other, or surrounding themselves with moats of landscaping. 

Planners on the other hand tend to work from regional and city-wide scales down to local neighbourhood levels, where their engagement with urbanism stops. But this approach can prove too crude to provide an accurate picture of what is going on at the important human scale. Planners should be able to analyse data to inform decisions - such as transport plans or changing land values -– at least down to the level of the individual street segment and ideally to the different buildings that make up the street level. 

An integrated modelling system should allow all of the buildings to talk to each other, then all the blocks in a neighbourhood to talk with each other, then all the neighbourhoods within a district to talk to each other and so on. This would be an 'Urban BIM': an Urban Building Information Management System that links different professional working across multiple spatial scales. 

But what does 'talking to each other' actually mean? It certainly involves visualisation of data on a common platform. But it also means going beyond visualisation into data analysis, correlation and modelling. Too many Smart City conference presentations and media articles focus on - sometimes obsess about - the visualisation of data, creating dazzling maps and movies, but often going little further. Captivating an audience is an important first step in raising awareness of the potentials of a data-driven approach; demonstrating that the data contains the intelligence to inform real urban changes is the next essential step. 

Urban modelling systems should therefore deliver the full potential of recent technological change by not only visualising data but, more importantly, identifying patterns within those data and establishing correlational relationships between different datasets. After all, urban planning and design practice necessitates the simultaneous analysis and resolution of multiple, complex issues. Urban modelling should seek to serve these ends, being able, for example, to associate input decisions on spatial layout and land use allocation to outcome phenomena such as land value, movement, crime risk and carbon emissions. In other words, a true Smart City solution should tell the “data story” from beginning to end. Urban modelling systems - based on algorithms that make sense of urban complexity - should equip architects, planners and stakeholders generally with "predictive analytics" that allow them to properly weigh up the pros and cons of different options (Reference 2). 

A SMART approach

 

A smarter approach is required to the creation of Smart Cities, one that lets users take a sequence of considered actions from initial data gathering to the final delivery of urban places. Space Syntax Limited, the architectural and urban design practice created at University College London in 1989, has developed the following 5-stage methodology: 

Sense/Survey 

First, capture useful 'urban performance' data such as the demographics of a particular place, the location of different types of retail, the types of employment and typical travel patterns as well as 'urban form' data including spatial accessibility, topography, building location, capacity and condition. 

Map 

Second, spatially visualise that data. For example: develop maps that geolocate the various urban performance and urban form characteristics. Use 2D and 3D techniques to explore the geolocated data. 

Analyse 

Third, use statistical tools to search the data for patterns, associations and correlations such as links between observed pedestrian movement data and spatial accessibility levels, or between spatial layout density and land value. Factor in the land use attraction created by shops and transport nodes. Identify algorithms that make sense of the data. 

React 

Fourth, create evidence-based policies, plans and detailed designs. use the visualised and analysed data to plan and design new places. Create scenarios of different future outcomes and consider the measures that need to be taken for any of these scenarios to occur. For example, if air quality is a key policy consideration then what switch from private to public transport, or from petrol to electric vehicles, would be needed? How would the city need to be redesigned to accommodate a new metro system, new bicycle infrastructure or new walking routes? 

Test 

Fifth, use algorithm-based predictive analytics to forecast the impacts of proposals in advance. For example, infer via a model simulation where residents are likely to want to travel to in the city and what sort of uptake there might be for a new bus route or cycle path. Use the results of these forecasts to discuss ideas with stakeholders, trying out different options for changing the area and reviewing how these would impact on the way the city works, in order to decide on which one would be most appropriate. 

Sense, Map, Analyse, React, Test = SMART 

Once a particular option has been decided on and implemented, the SMART process can be used to monitor how accurate the predictions were, in order to help refine and further develop the forecast models. In other words, the SMART process can be repeated through further sensing, mapping, analysis, reaction and testing. 

Space Syntax has followed the SMART process to pioneer the creation of urban modelling technologies for over 25 years, using the increasingly sophisticated capabilities of 2D and 3D digital software to produce ever more capable decision-making platforms. Much experience has been gained during this time to develop tools and methods that meet the needs of public, private and community organisations. The approach can be used in a strategic mode to robustly test the wide-scale urban planning impacts of various options, but are sufficiently detailed to inform architectural and landscape design discussions at a later point in the process. 

Experience has shown that systematic sensing, mapping and analysis of datasets provokes an informed, science-based reaction among policy-makers, planners and designers. In a similar way, stakeholders in consultation exercises respond more positively to objectively tested proposals than they do to instinctive, unsupported, opinion-based ideas. 

Space Syntax’s SMART approach to urban modelling equips users and their audiences with agile models of how urban areas work across a range of scales. These models are used to evaluate the likely impacts of planning and design proposals on local places as well as on the wider city. They work quickly to inform constructive discussions, which might otherwise degenerate into protracted and emotional conflict. This expedites evidence-based decisions, giving city authorities firm grounds, for example, to negotiate design changes with private developers. 

Key characteristics

 

Space Syntax Limited’s project experience is based on a 40-year fundamental research programme at University College London, led by Professor Bill Hillier (Reference 3). A close working relationship between design practice and university research has led to the application of deeply studied academic techniques in the planning, urban design and architecture. In return, the data and the learning from professional practice has been taken back into the rigorous research environment of a university setting. 

The two-way relationship between academic researchers and professional practitioners is the first fundamental characteristic of Space Syntax’s SMART approach: continuous co-creation for mutual benefit. 

The second key feature of Space Syntax’s methodology is its focus on human outcomes above data inputs. This people-first approach is fundamentally concerned with the social, economic and environmental impacts of the planning and design process on people. In practice, this means collecting data on human behaviour patterns: movement patterns on foot, bicycle and in vehicles; standing, sitting and interaction patterns in streets and public spaces; as well as data on socio-economic demographics, crime and health patterns. 

Third is a focus on space: Space Syntax has created an approach to spatial network analysis that uses the pattern of street connections as a “data framework” onto which other datasets are hung. This is done following the discovery that spatial networks have fundamental influences on human behaviour. More connected places, for example, can generate higher movement flows, land values and social potentials. Disconnected places can suffer from social isolation and economic underperformance. 

Space Syntax practitioners have found that it is essential to put spatial analysis of urban form at the 

heart of urban modelling, not least because it is in space that people meet each other and through space that they move. Social and economic transaction between people is, after all, the ultimate purpose of cities. Cities are intensifications of opportunities for people to form relations that drive the production of economic outputs, social networks and cultural phenomena. In order for these transactions to occur, people need to be able to move effectively from their place of origin to their place of interaction. 

Space Syntax research takes these two essential behaviours - transaction and movement - and places them centrally in its investigation of architectural and urban form. Space Syntax practice focuses on the design of public spaces, streets and movement systems for vehicles pedestrians and bicycles. Within buildings, Space Syntax examines the movement and interaction patterns of office workers, museum visitors, hospital staff and patients. 

Despite the introduction and proliferation of electronic communications in digital "space", it seems the importance of physical space has not diminished. People are moving to towns and cities in ever greater numbers. Nevertheless, an important interrelationship is emerging between human behaviour in the physical and digital domains, where people are increasingly moving and transacting seamlessly in one and the other, whether it is engaging in social media discussions while riding in a train or using augmented reality applications to enhance their awareness of the place they are in. People are meeting online and arranging to meet again in physical space. Future research and practice will need to address such 'digital urbanism'. 

Experience from stakeholder discussions shows that people understand the Space Syntax approach intuitively and respond positively to spatial analysis because, it seems, people “read” space naturally. The clear conviction is that, without spatial network analysis, urban models are missing their most essential ingredient. 

Modelling the impact of resource flows is a fourth key feature: the supply of energy and other utilities such as water and telecoms, as well as the handling of waste. Spatial network analysis is again powerful because many utilities are laid in line with street networks, both above and below ground. When urban planners set out a street pattern they are not only determining patterns of movement and human interaction but also the flows of resources that support those behaviours. 

The city is the largest intended object of collective human creation and its interactions with the natural environment need to be analysed and understood with modelling. This relationship is handled by the fifth fundamental characteristic of Space Syntax SMART models, which connects spatial network analysis to environmental features. For example, sun path analysis is linked to spatial accessibility analysis to modify the overall attraction of a highly connected street if that street is in full sun and a highly shaded, if less direct, alternative is available. In a similar way, linked analyses can be used to optimise the layout of an urban grid by aligning the geometry of the street pattern to the angle of the prevailing breeze. 

Together, these five fundamental characteristics define Space Syntax’s approach to the creation of 

Integrated Urban Models, models that help people make decisions to influence behaviour change towards more resilient, more sustainable ways of living. The Smart City process requires the capture and handling of data concerning people, urban form, resource flows and environmental factors: People, Urban form, Resource flows, Environmental factors = PURE 

 

By following a SMART process that integrates PURE data sources, an Integrated Urban Model is created. Space Syntax’s aim is to disseminate the Integrated Urban Model approach into general urban practice, thus sharing an approach to urban planning, design and governance that has been tried and tested for over a quarter of a century. Experience suggests that the demand for integrated urban modelling is growing as clients expect ever greater levels of evidence-informed planning and design. This was the case in the masterplanning of the City of Darwin as illustrated in the following case study. 

---

Case study: Darwin. Australia

 

In 2012 the City of Darwin secured a grant from the Commonwealth Government of Australia to produce a Masterplan for the City Centre of Darwin. The purpose of this plan is to provide a road map for development of the city for the next twenty years. To accommodate growth, the City Centre needs to be extended and reinforced. A key aim of the project was to facilitate private sector investment and population growth through the production of high quality public realm. 

Space Syntax was asked to join a design team led by Steve Thorne of Design Urban. Our role was to provide comprehensive urban analysis and strategic design input to support the creation of the Masterplan. 

Space Syntax provided an evidence-based platform for the creation of a spatial strategy. Our input had five work streams: 

Urban Data Collective

Data related to Urban Form (e.g. land use, transport and spatial form) and Urban Performance (e.g. pedestrian and vehicle movement, land value) were collected and analysed using Space Syntax software. 

Urban Performance Model

Associations between Urban Performance and Urban Form were analysed in order to provide an Urban Diagnosis from a number of perspectives; spatial, physical, economic and human behavioural. 

Spatial Design Strategy

Opportunities and constraints analysis was undertaken using the Urban Performance Model to develop a Spatial Design Strategy. This strategy, including as set of design objectives and principles, has guided the development of the Masterplan design. Option testing using Spatial Accessibility Models played an essential role in the design development process not only to assess design outcomes quantitatively but also to generate new design ideas and options. Frequent and direct communications with the design team, including attendance at workshops in Darwin were key to maintaining a dynamic design process and ensuring an integration between the baseline analysis and the development of the Spatial Design Strategy. 

Urban Forecast Model

Forecast models for pedestrian movement and for land value were constructed to quantitatively evaluate the impacts of the proposed designs. The outcomes provided useful indicators of the social and economic performance of the Masterplan. A Land Value Model, combining spatial layout and land use factors was used to identify $AUD 3.7 billion (RMB 17.25 billion) of land value increase resulting from the masterplan. 

Urban Design 3D Modelling

3D visualisation of the Masterplan in static and animated formats were created by Space Syntax and used in the stakeholder engagement process. Such visual representations of the Masterplan were effective in engaging the widest possible audience. 

The outcome

Space Syntax’s work has underpinned the creation of the Darwin City Centre Masterplan. Quantified indications of its social and economic impacts have facilitated meaningful discussions between different disciplines, bringing architects, transport engineers, planners and politicians together with a wide stakeholder base. 

Client 

City of Darwin, Northern Territory Government, Australian Government 

Partners 

Design Urban, Urbacity, Michels Warren Munday, Clouston Associates 

Year 

2013-14 

Summary

 

Cities can be smart in two ways. First, by harnessing technologies to improve the way that urban places are planned, designed, led and managed - the inputs. Second, by delivering buildings, places and systems that work more effectively for the people that use them - the outputs. This two-pronged approach applies to all aspects of Smart Cities. 

A Smart City approach should direct the capabilities of urban planners, designers and policy makers to: 

1. facilitate effective human transaction in new and existing places 

2. provide access to places of transaction, both physical and digital: on-land and on-line 

3. support the mobility required to access these places of transaction by providing networks of connectivity for all modes of transport, both physical (walking cycling rolling driving) and digital 

4. take an outcomes-oriented (ie transactions & emissions) approach first and foremost, aware of the inputs required (ie materials, energy & mobility) to achieve these desired objectives 

5. employ effective analytic and forecasting tools aimed at social economic and environmental impacts. 

As a pioneer of urban technologies, Space Syntax has developed a unique approach to Smart Cities, with a particular emphasis on spatial layout analysis and data integration. Its Integrated Urban Modelling platform has proven to be an effective means of creating places that improve resource efficiency and benefit environmental impacts. Built on open source software and disseminated via open access websites, the Space Syntax approach is intended to be as available and accessible as possible. In this way, the growing demand for "smart" urban inputs can be addressed and the global imperative to deliver "smart" urban outcomes can be met. 

REFERENCES 

1. Future Cities Catapult, "Future Cities: UK capabilities for urban innovation", 2014, p35 

2. Data and Analytics Task Group, UK Government Smart Cities Forum, "Data and Analytics: Resources for Smart Cities", UK Government 2015, p4 

3. Hillier, Bill, "Space is the machine: a configurational theory of architecture", CreateSpace publishing 2015, www.spaceisthemachine.com