Is It Possible to Predict a City’s Life Based on the Shape of Its Neighbourhoods?

© HSE University / Andrey Morkushin

Is it possible to predict, based on the configuration of streets and buildings, where a café will open or where traffic congestion will occur? Participants in the Spatial Analysis and Modelling of Urban Processes research and study group use open data and machine learning to identify universal patterns. Alexander Sheludkov and Eduard Somov discuss the purpose of comparing cities, the need for new forms of urban statistics, and how open data is transforming approaches to urban studies.

— How did the idea for your project emerge? Does it have a backstory?

Alexander Sheludkov, Associate Professor at the Joint Department with the RAS Institute of Geography, HSE Faculty of Geography and Geoinformation Technology, and Programme Track Supervisor of the Spatial Data and Applied Geoanalytics Master's Programme

— The question of how the physical form of cities—the configuration and spatial arrangement of buildings, streets, and neighbourhoods—relates to the economic and social development of urban areas, including the density of economic activity, the location of services, and the distribution of pedestrian and vehicle flows, is one of the fundamental issues in geography, urban economics, and architecture.

Every city is unique, yet the patterns of urban development are remarkably similar. Every city has a centre and peripheral districts, busy pedestrian streets lined with shops and cafés, major highways and quiet secluded streets. There are affluent neighbourhoods and poorer areas; places that bustle with activity during the day and fall quiet in the evening, and others where life is concentrated at night. The physical fabric of a city, its morphology, creates the conditions for these differences to emerge, but it can also change under the influence of economic forces and social interactions, leading to the redevelopment of buildings and entire districts.

Morphometry and spatial statistics seek to translate these differences into numbers—to describe mathematically how the configuration of streets and buildings relates to specific manifestations of human activity in urban space.

The problem is that the models describing this relationship were developed at a time when accurate, detailed data for a large number of cities was not available. Today, such data—particularly that describing urban morphology—has become accessible

There are, for example, open-source mapping services such as OpenStreetMap. The capabilities of automated data processing, including machine learning algorithms and methods for detecting nonlinear relationships, have expanded significantly. Together, these two factors open new avenues for research into whether the relationships between urban morphology and urban functions follow universal patterns and, most importantly, whether morphological data can be used to predict the future economic and social dynamics of cities.

Eduard Somov, Associate Professor at the Faculty of Geography and Geoinformation Technology, Programme Track Supervisor of the Spatial Data and Applied Geoanalytics Master's Programme

— We drew inspiration from several sources. One example is the work of European researchers on the taxonomy of urban forms. Their studies also rely on open data and scalable methods; however, their findings cannot be directly transferred to Russia, where urban development trajectories differ significantly from those in Europe. Another source of inspiration is the analytical report by colleagues from the Faculty of Urban and Regional Development, Urban Sprawl Planning (2021). It is the first large-scale study of urban growth processes in Russia, based on a sample of more than 90 cities. We also plan to work with a large dataset, but our focus is on spatial data.

— What is the goal of your project?

Alexander Sheludkov 

— The ultimate goal of the project is to develop alternative urban statistics based on the morphometric characteristics of individual buildings, streets, and neighbourhoods, while also capturing how the city functions as a whole: how accessibility varies across areas where jobs and social activity are concentrated, which routes residents take, and, ultimately, how cities differ in the way they operate. We aim to move towards scalable approaches that rely entirely on open data.

— What is the relevance of your project?

Alexander Sheludkov 

— Geography has long since evolved into a science of big data, particularly in the field of satellite image processing. Thousands of studies are conducted worldwide each year based on such datasets. Urban studies, however, lag behind in this respect: contemporary urban research is increasingly data-intensive, yet detailed statistics at the level of cities and neighbourhoods are either unavailable, fragmented, or too costly and complex to collect. There is a lack of standardised information and widely accepted methods for processing it, which limits the possibility of scalable research across large samples. Through our work, we aim to help close this gap, at least in part.

We are confident that such data and methods will be in high demand both in academia and beyond

— What results do you expect to achieve?

Alexander Sheludkov

— The first year of the project is devoted to working with individual urban polygons, where we test hypotheses and refine our tools. Yaroslavl, Tyumen, and Vladivostok have been selected as case studies. These are three large cities located in different parts of Russia that differ significantly in many respects: their time of foundation, topography (which strongly influences urban morphology), patterns of development, and rates of growth. By selecting such diverse cities, we aim to capture a wide range of urban forms.

Eduard Somov

— One of our main objectives is not simply to obtain sets of figures for each city but to deepen our understanding of how these indicators relate to the functioning and development of urban districts.

By functional development, we primarily mean the mix and saturation of functions (such as shops, cafés, and other services) present in a given area, such as a street segment, neighbourhood, or district. By mobility patterns, we refer to the intensity of movement and how it is distributed over time.

However, even in the absence of such patterns, the resulting set of indicators and the database itself, providing a detailed description of various aspects and parameters of the urban spatial structure, will constitute a significant outcome of the project.

Such a database could be widely used in a broad range of urban studies

— What challenges have you encountered, or do you expect to encounter, during the implementation of your project?

Alexander Sheludkov

— There are complex methodological challenges. Any model is a simplification of reality, and none of our models can capture all aspects of urban development or internal structure. In any case, we will only obtain an approximation of reality. In addition, as the volume of data and the number of variables increase, the potential for meaningful interpretation decreases. There are hundreds of metrics that describe urban morphology. How do we choose the most relevant ones? How should they be combined, and what exactly do they indicate? We need to strike a balance between the depth and breadth of the analysis, without losing our understanding of the underlying processes.

Eduard Somov

— One of the challenges we will have to address is scalability. Although we carefully selected pilot cities and identified sites with very different characteristics, they do not cover the full range of possible cases. This means that we will need to carry out additional iterations and make adjustments to the methodological part of the project.

— How were the project participants selected? Could you tell us about the key members of the team?

Alexander Sheludkov

— The project participants are primarily students from the Faculty of Geography and Geoinformation Technology.

Each of them develops their own research area through term paper projects that align with the objectives of the research and study group

My personal research interests focus on modelling the mobility of urban residents, and the project opens up new opportunities to study these processes.

Eduard Somov co-supervises the project with me. He has extensive experience in applied research on urban areas, particularly in analysing and optimising the placement of commercial and other infrastructure facilities.

Oleg Kiselev, a graduate of the Faculty of Urban and Regional Development and a visiting lecturer at our faculty, is also part of the team. His professional work focuses on the development of new forms of urban transport, in particular the study of how individual mobility influences urban life. It was he who suggested shifting the project’s focus towards urban morphology.

— What are the prospects for the project's further development?

Eduard Somov

— Ultimately, we aim to develop a universal tool for assessing the urban environment based on methods for analysing urban morphology and spatial statistics. A key advantage of this tool will be its ability to operate using, on one hand, publicly available data and, on the other, highly detailed and comprehensive datasets. We believe this will make the tool particularly valuable in terms of optimal balance between the resources required for its use and the insights it can deliver.