Chair: Janusz Dygaszewicz/Julien Gaffuri
Room: S4B Lajkonik
Time: 11:30 - 13:00
Date: 28 June
Session 40 - papers & presentations
|Title: <<< Integration of GEOSPATIAL DATA within the statistical production process - GEOINQ >>>
The use of cartography has supported census data collection at Statistics Portugal since 1981.
Since 2006, with the production of the enumeration areas dataset to support the 2011 census, Statistics Portugal has been developing a Spatial Data Infrastructure (SDI) that is currently being used, in a transversal way, to promote the integration of the geospatial data in the statistical production process, in order to achieve efficiency and accuracy, within sampling process, data collection or dissemination. In 2011 Statistics Portugal built a national geodatabase comprising all the georeferenced buildings from the 2011 Census (BGE). BGE is a point based coverage that is being continuously edited in a internal quality control process and updated by the Municipalities which provide Statistics Portugal, on a monthly basis, all the completed buildings and buildings permits including X,Y location and addresses. After the 2011 Census, Statistics Portugal evaluated the possibility of implementing a geographic tool that would allow the visualization of buildings and dwellings units of survey samples. For this purpose, GeoINQ was developed to allow the visualization of the location of sample buildings and to provide management and control functionalities of the data collection process. In addition, GeoINQ has editing tools to update the BGE by field interviewers.
It is a Geographic Information Systems (GIS) WEB solution developed in order to integrate geospatial data into the production process of official statistics in an innovative way. It allows greater efficiency and rationalization of the resources especially in the household’s surveys by supporting the data collection process. GeoINQ is integrated, via webservices, with Statistics Portugal Global Survey Management System (SIGINQ-IE), and consumes a set of services and geographic data of subjects of the INSPIRE directive.
|Title: <<< Improving the quality with spatial sampling >>>
Over the last years, many initiatives have been undertaken to help the National Statistical Institutes construct a fully geocoded information system. Such a point-based system is precisely the starting point of a handbook of spatial statistics being written by Insee and to be released by the end of June 2018. The latter,funded by Eurostat,will drawa list of statistical methods that rely on theavailabiltyof the (x,y) coordinates of the statistical units. Ranging frommeasures of spatial autocorrelation to spatial econometricsfor panel data, these methods might be helpful to improve the production, the dissemination or the analysis of statistical results. The issue of spatial sampling fully falls within the scope of the handbook, and as such will be more precisely dealt. The presentation at the conference will focus on the issue of spatial sampling. It aims at proving that a geocoded sampling frame might help better carry out surveys. On the one hand, knowing the position of the statistical units help better organise the field work for face-to-face surveys. This can be done, for instance, thanks to Primary Units having very good spatial features. On the other hand, at the selection level, the sampling design might better spread the selected units over the territory. This strategy can be efficient to improve the precision estimation for high spatial correlated variables.
|Title: <<< Making census statistics more relevant – towards geo-enabled statistics >>>
At a time when there is evidence of public distrust of statistics and of distancing between policy makers and the public, it is important that official data should relate closely to the realities of people's lives. One way to do this is to provide geographically detailed statistics that describe the situation at local level. In many contexts, statistics at the level of towns and neighbourhoods are the most relevant and most meaningful to the citizens and local authorities to support the local level policy decisions that are of importance to people's day-to-day lives. For the 2021EUcensus, building on the success of the GEOSTAT2011 grid and in response to growing needs for high resolution population statistics, the work is advanced on the 13 key census topics geocoded to a 1km² grid. It will be a new departure, allowing new types of statistical and spatial analyses for areas that unlike the traditional NUTS areas can be flexibly defined according to the needs of policy makers and researchers. Grids as a time series will also assure geographical comparability over time. One use of such data to ensure the efficient allocation of funds (e.g., European Cohesion Policy €50 billion/year) is the analysis of populations for small urban and rural areas, and assessing the accessibility of services such as education and health for citizens. This paper presents the current plans for developing the grid data collection for the 2021-census and discusses the main legal, organisational and methodological issuesto be addressed. It describes the innovative and complex development, requiring a separate legal act and a number of solutions to emerging technical and methodological issues, e.g.Statistical Disclosure Control. Itlooks forward towards the post-2021-EUcensuses as the attractiveness of providing population data to a grid made it one of the focal points of the developments.
|Title: <<< Towards better quality of statistical geospatial data production - harmonization of statistical and geodetic divisions in the context of the 10 Level Model >>>
According to adopted rules, in Poland boundaries of statistical units are aligned with boundaries of cadastral units where applicable. In connection with the need to preserve the limits of housing and people in the statistical units and taking into account the diversity of terrain and population density, statistical regions and census enumeration areas in rural and urban areas have different extend. To preserve collinearity in case of any changes made by the cadastral service appropriate changes must be made on the statistical side. In reference to keeping a consistency of both divisions Polish proposal of the 10 Level Model will be used to better understand and develop statistical and geodetic reference framework. The proposed model should be the subject of intensive works in order to overcome existing barriers and as a starting point to make practical progress in the methodology of combining spatial data with statistical data. Development of statistical division based on geodetic division should provide better interoperability of sets of data and raise possibilities of statistical geospatial analyses. However such harmonization will cause that quality of statistical geospatial data as well final statistical products and conducted analyses will depend on the quality of input spatial data from external administrative registers. That is why assessment of the overall quality of external spatial data sets, and especially the quality of data which they include is essential. Recently Polish official statistics worked on the project which aim was the improvement of the use of administrative sources. As a result, on the basis of Polish experience, the methodology of assessing the usability of administrative data sources including spatial data registers has been elaborated.
|Title: <<< Accessibility statistics and data integration: from remoteness and public transport to cultural accessibility >>>
Accessibility research has been a relevant part of today’s geographic information science. Data sources of official statistics offer plenty of relevant administrative data of the population itself and of many kinds of services it is potentially using. These are also combined to statistical products for customers of Statistics Finland. This paper discusses many aspects of accessibility statistics. The target here is the whole country. First, the remoteness (index) estimation that has been conventionally computed by Euclidean buffer populations for the Ministry of Finance. Studies funded by the ministry clearly promote the use of a road network based estimation. Elementary school accessibility has been taken into account as well. These kinds of statistics at the municipal level offer valuable information for the relevant authorities. Recently, the need for cultural accessibility statistics has been raised at Statistics Finland. This requires data from many sources: libraries, theatres, movie theatres, orchestras, museums, festivals, etc. Naturally, the road network application is much more suitable than linear distances. The most advanced application is commuting statistics. Here, the travel time itself becomes relevant and estimating that during the rush hour is a complicated task requiring data integration of many different sources. On the other hand, new data sources such as public transport web services enrich the modelling much. Greener commuting, bicycling, has received the most attention, however. This paper both informs and motivates for discussions of current and future research of statistical accessibility.