Geospatial analysis at Eurostat
Data from April 2015
This article presents spatial analysis, a set of techniques and tools to study spatio-temporal relationships inherent in data. The article further sets out the possibilities for its usage and gives some examples where spatial analysis has been conducted with statistical data of the European Union (EU). GISCO, the team managing the 'Geographical information system of the Commission', is a centre of excellence of spatial analysis within the European Commission.
What is spatial analysis?
The term spatial analysis defines a set of techniques used to visualise, process or analyse data with spatial attributes. This definition implies a very wide-reaching field for the spatial analysis, as much data are inherently spatial, being referenced to locations in two- or three-dimensional space. Collecting the spatial attributes (geographic location) of statistical or other data is what creates the spatial data and hence allows the spatial analysis.
The term spatial analysis is often used in different contexts:
- A map or a time sequence of maps may be considered as spatial analysis when visualising the spatial aspects of datasets according to chosen criteria (see Map 2).
- Processing and analysing spatial data using the set of functions provided by many geographical information system (GIS) software packages such as overlaying, buffering, filtering, distance computations, etc. Descriptive statistics — calculating maxima, minima, means, variances and others — are also referred to as ‘spatial analysis’ (see Map 1).
- A more advanced usage of the term refers to the combining of spatial and statistical techniques to analyse spatial and temporal data in a number of scientific fields such as environmental, agricultural, medical, social, economic and political sciences, as well as urban planning, engineering and other sciences. This type of spatial analysis typically requires access to comprehensive stack of geospatial, statistical and other thematic data sources and is normally not conducted by statistical offices.
Map 1, for example, shows the number of airports the resident population around the airport of Burgos in Spain can reach within a 2-hour journey. This analysis was an element in a recent performance audit carried out by the European Court of Auditors (ECA) which investigated the return-on-investment of public spending on airport infrastructures.
Spatial analysis for statistical purposes
Statisticians seldom discuss spatial analysis, although spatial analysis techniques are widely used to create the reference for the collection of statistical data and are involved in much of the collection, dissemination and analysis of statistical data. The following examples will illustrate how spatial statistics and spatial analysis are used in official European statistics.
Nomenclature of territorial units for statistics
Large amounts of European Union (EU) regional statistical data are collected, harmonised but also (geo)spatially located using the nomenclature of territorial units for statistics (NUTS) and local administrative units (LAUs) as references. The collected data are used for the socio-economic analysis of the regions, applying spatial analysis techniques and references to EU regional policies. The NUTS and LAU classifications and geospatial statistical datasets are therefore the starting point for any EU-wide spatial analysis. The maintenance of this hierarchical system of administrative units and the respective geographical information (GI) datasets is a complex undertaking involving demographic and economic indicators to outline the boundaries of the statistical units.
Regional statistics are a helpful tool to understand the regional diversity. Considering national figures alone does not reveal the full picture of what is happening in the European Union; indeed, there are often significant differences between regions of the same country when one looks at smaller geographical areas. Regional statistics are based on a harmonised convention in the definition of regions which is contained in the NUTS. Identifying, mapping and analysing regional differences, using NUTS and LAU classified statistical data, provide an opportunity for interesting insights based on (geo)spatial variation of diverse statistical indicators.
The Eurostat regional yearbook provides an overview of official, regional statistics that are available within Europe. See Regional Yearbook foreword. The Eurostat Statistical Atlas is an online application for the dynamic visualisation of the maps from the Eurostat regional yearbook. Regional data may be easily compared against other statistical data for the same administrative unit, against the data for the adjacent units, displaying infrastructure and geographical features.
What is the boundary of a city?
The need for spatial analysis becomes evident when the question ‘What is a city?’ is asked. Any analysis of the spatial dimension of EU cities starts with this question, which then leads to others, focusing on the spatial characteristics of the cities. What is actually the ‘core city’ and what is its area? What is the boundary of ‘the functional urban area’ (FUA) and how can these boundaries be identified? How can the city be divided into ‘sub-city districts’ (SCD) applying population criteria?
The City data collection system provides information on different aspects of the quality of urban life in a cross-section of Europe’s cities’. The collected data is used for analysis of the spatial dimensions of the European cities, focusing on development of the city infrastructure and socio-economic dynamics. Regional and city development is observed from the perspective of urban-rural connectivity, administrative and political boundaries, interactions of the city with its surroundings and other phenomena with spatial dimensions.
Degree of urbanisation classification (DEGURBA)
The degree of urbanisation (DEGURBA) is a classification of LAU2s created using spatial analysis techniques. The classification was developed jointly by DG AGRI and DG REGIO with the support of the Joint Research Centre (JRC) and Eurostat. Local administrative units level 2 (LAU2) are classified into three types of area based on the share of local population living in urban clusters and in urban centres:
- densely populated area (cities/large urban area);
- intermediate density area (towns and suburbs/small urban area); and
- thinly populated area (rural area).
Urban, suburban and rural areas are concepts used by policymakers, researchers, national administrations and international organisations. In the EU Member States and the European Free Trade Association (EFTA) countries, the degree of urbanisation helps to structure data according to this classification from a wide range of surveys, including the Labour Force Survey (LFS) and the Survey on Income and Living Conditions (SILC). See also Working Paper on 'Harmonised definition of cities and rural areas']
Data by degree of urbanisation is presently available for the following statistical domains:
- labour market;
- education and living conditions; and
- welfare and tourism.