European cities - spatial dimension
From Statistics Explained
This article takes a closer look at the spatial dimension of cities in the European Union (EU). Cities are often analysed as distinct, unconnected dots on a map, but this fails to represent the linkages between a city and its hinterland and between different cities. Recent developments in transport, communication and information technology infrastructure have considerably increased the flow of people and resources from one area to another and, thus, have rendered the spatial dimension even more important. Urban–rural connectivity and inter-urban relations have become critical for balanced regional development.
Data on European cities are being collected in the Urban audit project. The project, whose ultimate goal is to assist cities to improve the quality of urban life, supports the exchange of experiences among European cities, helps to identify best practices, facilitates benchmarking at European level, and provides information on the dynamics of urban life both within the cities and between them and their surroundings.
Discovering the spatial dimension
Cities are usually displayed as distinct unconnected dots on a map. This visualisation method increases visibility but it misrepresents reality and distorts the understanding of linkages between a city and its hinterland and the understanding of linkages between cities. Cities can no longer be treated as discrete unrelated entities without a spatial dimension. Recent developments in transport and information and communication technology infrastructure have eased the flow of people and resources from one area to another considerably. Urban–rural connectivity and inter-urban relations have become critical for balanced regional development.
To facilitate the analysis of the interaction between the city and its surroundings for each participating city, different spatial levels were defined. Most of the data are collected at core city level, i.e. the city as defined by its administrative and/or political boundaries. In addition, a level called the larger urban zone was described. The larger urban zone is an approximation of the functional urban area extending beyond the core city.
Map 1 illustrates the cities participating in the Urban audit data collection, showing the boundaries of core cities and larger urban zones. Not surprisingly, the largest cities in Europe in terms of population — London, Paris, Berlin and Madrid — tend to have the greatest larger urban zones in terms of area, and are readily identifiable on the map. In most cases, the larger urban zone includes only one core city. However, there are exceptions, such as the German Ruhr area, which includes several core cities (see inset in Map 1). The demarcation of core cities is illustrated in detail in Map 2 while the larger urban zones are shown in Map 3. The spatial data used to produce most of the maps presented in this article are available from the Geographic Information System of the European Commission (GISCO) — a permanent service of Eurostat.
Throughout Europe’s history — in ancient Greece, in ancient Rome, and in the Middle Ages — a city represented as much a political entity as a collection of buildings. And this collection of buildings was usually surrounded by fortified walls. As the city grew, the walls were expanded. In the modern era, the significance of the city walls as part of the defence system declined and most of them were demolished. The boundary of the city as a political entity and the boundary of the built-up area were no longer linked and the location of these boundaries is no longer clearly evident on the ground. Nowadays, a city could be designated as an urban settlement or as a legal, administrative entity. The Urban audit uses this later concept and defines the core city by political boundaries. This ensures that data are directly relevant to policy-makers.
Map 2 illustrates the difference between the two concepts using the examples of Hamburg (Germany) and Lyon (France). Maps in the top row show the land cover based on Corine Land Cover 2000 (CLC2000) data in the area surrounding the cities. Different land covers were grouped into 44 classes in the CLC2000 system. Each colour on the map represents a different land cover class. Some of these classes are particularly important for our analysis of cities. Red areas, for instance, are territories covered with urban fabric, such as roads, residential buildings, buildings belonging to the local administration or to public services, etc.
Purple areas are used for commercial or industrial purposes. Light purple represents green urban areas like parks, botanical gardens, etc. The areas of these three land cover classes lying less than 200m apart were merged together to define ‘built-up’ area. Port areas, airports and sport facilities were included, if they were neighbours of the previously defined ‘built-up’ area.
As a next step, road and rail networks and water courses were added if they were within 300 m of the area defined beforehand. The area identified by this procedure is called the ‘urban morphological zone’ (UMZ). The urban morphological zones of Hamburg and Lyon are shown in the middle row of Map 2. These maps also make it possible to compare the UMZ and core city in terms of area. In Hamburg 82 %, and in Lyon 73 %, of the area of the UMZ lies within the boundaries of the core city. In terms of population, the overlap is even greater: 90 % of the population of the core city of Hamburg lives in the UMZ, and in Lyon, the respective figure is 98 %. As we expected, the two areas are not identical but they overlap each other to a large extent, thus ensuring that the data collected at core city level are relevant and meaningful for the morphological city as well.
To measure spatial inequalities within the city, the area of the core city was divided into sub-city districts. Sub-city districts were defined in such a way as to keep to the population thresholds set — between a minimum 5 000 and a maximum 40 000 inhabitants — as far as possible. The bottom row of Map 2 illustrates the sub-city districts of Hamburg and Lyon. Key demographic and social indicators are available in the Urban audit database for the more than 6 000 sub-city districts.
Larger urban zones
City walls, even if they are preserved, no longer function as barriers between the people living inside and outside the city. Students, workers and people looking for healthcare or for cultural facilities regularly commute between the city and the surrounding area. Economic activity, transport flows and air pollution clearly cross the administrative boundaries of a city as well. Consequently, collecting data exclusively at core city level is insufficient. It is commonly agreed that we have to widen our territorial perspective. However, the way to measure how far the functional influences of a city go beyond its immediate boundaries varies.
Map 3 uses the examples of Barcelona (Spain) and Zagreb (Croatia) to illustrate how the functional urban area was defined in the Urban audit. Maps in the top row are similar to the top row of Map 2 portraying the land cover of the selected area. The larger urban zone around the core city tends to be more ‘green’, both on the map and also in real terms. Areas covered with forests and shrubs are coloured green on the map. Yellow and orange indicate areas in agricultural use, such as arable land and fruit trees. As a first step to defining the larger urban zones, we looked at the number of people commuting from municipalities to the core city. The middle row of Map 3 displays the different commuting rates. A commuting rate of 10 % means that one in 10 residents living in the municipality commutes to work in the core city. As we can see on the map, large cities, such as Barcelona and Zagreb, attract people living up to 100 kilometres away to work in the city. As a second step, a threshold was set for looking at the commuting pattern. Municipalities above this threshold were to be included but ones below not. Given the different national and regional characteristics, different thresholds were used within the range of 10–20 %. Finally, the list of municipalities to be included in the larger urban zone was revised to ensure spatial contiguity (proximity) and data availability. By definition, the larger urban zone always includes the entire core city. The boundaries of the larger urban zone of Barcelona and Zagreb are displayed in the bottom row of Map 3.
This demarcation process was used in most participating countries, but there were also exceptions and departures from this which limit the overall comparability of the larger urban zones to some extent. That said, defining a perfect functional urban area — based on a perfectly harmonised methodology across Europe for which no statistical information is available — would be completely in vain. Graphs 1 and 2 compare the different spatial levels used for European capitals in terms of population and area. In Bucuresti (Romania) more than 80 % of the larger urban zone population lives within the core city. At the other extreme, in Luxembourg (Luxembourg), less than 20 % of the larger urban zone population lives within the core city. This low percentage suggests that the core city of Luxembourg is slightly under-bounded. This means that a considerable share of the urban population lives outside the administrative city limits.
For very under-bounded capitals — like Paris (France) or Lisboa (Portugal) — an additional spatial level, the ‘kernel’, was introduced. The kernel is an approximation of the built-up area around the core city. The only exception is London (United Kingdom), where the kernel was defined to match the core city of Paris in terms of population to make for easier comparison between the two largest cities in Europe. In terms of area, the picture is more uniform, as for the majority of capitals, the core city makes up less than 20 % of the area of the larger urban zone.
So far, we have seen that larger urban zones tend to have a lower population density and a higher percentage of green areas than core cities. Using the indicators calculated in the Urban audit, we can analyze the demographic, economic, environmental, social and cultural characteristics (similarities and differences) of the two spatial levels. To illustrate this, Graph 3 compares the travel to work patterns in terms of mode of transport in selected capitals at different levels. The inner circles of the pie charts show the modal split in the core city. In the core city of København (Denmark), for example, the majority of people ride their bikes to work, 30 % of them use public transport and 25 % travel by car. The outer circle shows the share of transport modes in the larger urban zone. As expected, the proportion of journeys to work by car is consistently higher in the larger urban zone than in the core city, with the sole exception of Bratislava.
Where do families settle? Where do companies locate? Where do tourists stay? In the core city or in the area of the larger urban zone outside of the core city? We encourage readers to probe deeper into the Urban audit database and to explore the indicators depicting the spatial dimension to discover answers to these questions and many more.
The book entitled 'The spatial economy' , co-authored by Paul Krugman, winner of the 2008 Nobel Memorial Prize in Economic Sciences, states: "Agglomeration […] occurs at many levels, from the local shopping districts that serve residential areas within cities to specialised economic regions like Silicon Valley or the City of London that serve the world market as a whole. […] Yet although agglomeration is a clearly powerful force, it is not all-powerful: London is big, but most Britons live elsewhere, in a system of cities with widely varying sizes and roles. It should not, in other words, be hard to convince economists that economic geography […] is both an interesting and important subject."
In this article, we have focused on the various spatial levels used in the Urban audit. These provide a platform for analyzing the dramatically uneven distribution of population across the landscape and the agglomeration at district, at city and at regional level. Our intention was to convince readers that ‘statistical geography’ is both an interesting and an important subject.
The Urban audit celebrated its 10th anniversary in 2009. The ‘Urban audit pilot project’ was the first attempt to collect comparable indicators on European cities, and was first conducted by the European Commission in June 1999. The past decade has brought many changes, and constant efforts have been made to improve the quality of the data — including coverage, comparability and relevance.
So, where are we now? The list of indicators has been enhanced to take account of new policy needs, while the periodicity of the audit has been reduced to satisfy users, and geographical coverage has been extended following successive rounds of EU enlargement.
Enhanced list of indicators
There have been three major revisions of the list of indicators so far. Policy relevance, data availability and experience with previous collections have been reviewed to produce the current list of more than 300 indicators. These indicators cover several aspects of quality of life including:
- labour market;
- income disparity;
- local administration;
- educational qualifications;
- the environment;
- travel patterns;
- the information society;
- cultural infrastructure.
They are derived from the variables collected by the European Statistical System. Data availability differs from domain to domain: when it comes to demography, for example, data are available for more than 90 % of the cities, whereas for the environment, data are available for fewer than half of the cities. In 2009, we are due to introduce new indicators to symbolise the relationship between the city and its hinterland.
Moving from five-year periodicity to annual data collection
Four reference years have been defined so far for the Urban audit: 1991, 1996, 2001 and 2004. For the years 1991 and 1996, data were collected retrospectively only for a reduced number of 80 variables. Where data for these years were not available, data from adjacent years were also accepted. In 2009, Eurostat launched an annual Urban audit, requesting data for a limited number of variables. The annual data will help users to monitor certain urban developments more closely.
Extended geographical coverage
The pilot study in 1999 covered 58 cities from 15 countries. Since then, the number of participating countries has doubled and the number of cities has grown six-fold. At present, the Urban audit covers 362 cities from 31 countries — including the EU, Croatia, Turkey, Norway and Switzerland. The 321 Urban audit cities in the EU have more than 120 million inhabitants, covering approximately 25 % of the total population. This extended sample ensures that the results give a reliable portrait of urban Europe.
The number of cities was limited and the ones selected should reflect the geographical cross-section of each country. Consequently, in a few countries, some large cities (over 100 000 inhabitants) were not included. To complement the Urban audit data collection in this respect, the Large city audit was launched. The Large city audit includes all ‘non-Urban audit cities’ with more than 100 000 inhabitants in the EU. For these cities, a reduced set of 50 variables is collected.
We invite all readers to explore the wealth of information gathered in the past 10 years by browsing the Urban audit data on Eurostat’s website.
Improving the attractiveness of regions and cities is one of the priorities targeted by the renewed Lisbon Strategy and the EU's strategic guidelines on cohesion policy for 2007–2013. Quality of life is crucial in attracting and retaining a skilled labour force, businesses, students, tourists and, most of all, residents to a city. Assessing the current situation is a prerequisite for any improvement, development and future monitoring. The Urban audit is a response to this demand for assessment. This data collection provides information on the different aspects of the quality of urban life in Europe’s cities.
The Urban audit has become a core task of Eurostat. Even so, the project would not have been possible without sustained help and support from the cities themselves, the national statistical institutes and the European Commission's Directorate-General for Regional Policy.
What makes the Urban audit unique?
The Urban audit exercise can now look back over almost a decade of trials, errors, and achievements. Several concepts were tested and large volumes of data were collected during the pilot study in 1999, the first large-scale data collection round of 2003/04 and the most recent collection round of 2006/07. The data which passed the quality control procedures have, since April 2008, been available in Eurostat’s statistical databases. The uniqueness of the Urban audit data set lies in the extent of its three main dimensions: its wide choice of indicators, its large geographical coverage, and its decade-long time series.
What makes a city attractive? Residents are attracted to cities with a high quality of life, businesses are attracted to cities with a good skills base and infrastructure, students are attracted to cities with a good university or college, and tourists are attracted to cities with cultural value and mild weather, etc. As a result, a city’s attractiveness is determined by a number of factors. In this article, we mentioned a few, such as demographic characteristics, economic structure, the environment and social aspects. However, several other elements could be analyzed. We encourage readers to probe deeper into the Urban audit database and discover which cities they find attractive.
Further Eurostat information
- The Urban Audit — measuring the quality of life in European cities - Statistics in Focus 82/2008
- Urban audit (t_urb), see:
- Demographic indicators - Total resident population (tgs00013)
- Demographic indicators - Households with children aged 0-17 (tgs00014)
- Social indicators - Average living area in square metres per person (tgs00015)
- Social indicators - Mortality rate for <65 from heart diseases and respiratory illness (tgs00016)
- Social indicators - Number of car thefts per 1 000 population (tgs00017)
- Training and Education indicators - Students in higher education (ISCED 97 levels 5-6) per 1 000 population (tgs00018)
- Economic indicators - Activity rate (tgs00019)
- Economic indicators - Unemployment rate (tgs00020)
- Travel and transport indicators - Number of registered cars per 1000 population (tgs00021)
- Environment indicators - Collected solid waste per capita per year (tgs00022)
- Cultural indicators - Total annual tourist overnight stays in registered accommodation per year (tgs00023)
- Urban audit (urb), see:
- Key indicators for core cities (urb_ikey)
- Derived indicators for core city (urb_icity)
- Derived indicators for larger urban zones (urb_iluz)
- Derived indicators for sub-city districts (urb_iscd)
- Reduced set of derived indicators for 570 cities (urb_ilca)
- Data collected for core city (urb_vcity)
- Data collected for larger urban zones (urb_vluz)
- Reduced set of data collected for 570 cities (urb_vlca)
- Perception survey results (urb_percep)
- Regions and cities, see:
- City statistics - Urban Audit
- European cities - demographic challenges
- Statistics on European cities
- Urban-rural typology (background article)
- ↑ Masahisa Fujita, Paul R Krugman and Anthony Venables, The spatial economy: cities, regions and international trade. MIT Press, 2001.