Skip to Main content Skip to Navigation

Immersing evolving geographic divisions in the semantic Web

Abstract : Nowadays, the volume of data coming from the public sector is growing rapidly on the Open Data Web. Most of data come from governmental agencies such as Statistical and Mapping Agencies. Together, these public institutions publish geo-coded statistics that are of utmost importance for policy-makers to conduct various analyses upon their jurisdiction, in time and space. However, through times, all over the world, the subdivisions of such ju- risdictions (portions of space on Earth) delimited by or, under the control of human groups (e.g., administrative or electoral areas) are subject to change: their names, belonging or boundaries change for political or administrative reasons. Likewise, the Territorial Statistical Nomenclatures (TSNs) that are sets of artifact areas (although they usually correspond to political or administrative structures) built by Statistical Agencies to observe a territory at several levels (e.g., regions, districts, sub-districts) also change over time. Changes in TSNs are an obstacle to maintain the comparability of socio-economic data over time, unless past data are recalculated according to present geographic areas, a complicated process that, in the end, hide the territorial changes. Then, territorial changes lead to breaks in the statistical series, and are sources of misinterpretations of statistics, or statistical bias when not properly documented. Therefore, solutions for representing different versions of TSNs, and their evolution on the Open Data Web are to be proposed in order to enhance the understanding of territorial dynamics.In this thesis, we present the Theseus Framework with reference to philosophical issue raised by the Ship of Theseus that, according to legend, was rebuilt entirely over the years, every plank of the ship being replaced one by one. This software framework adopts Semantic Web technologies and Linked Open Data (LOD) representation for the description of the TSNs’ areas, and of their changes: this guaranties the syntactic and, moreover, semantic interoperability between systems exchanging TSN information. Theseus is composed of a set of modules to handle the whole TSN data life cycle on the LOD Web: from the modeling of geographic areas and of their changes, to the exploitation of these descriptions on the LOD Web. All the software modules rely on two ontologies, TSN Ontology and TSN-Change Ontology, we have designed for an unambiguous description of the areas in time and space, and for the description of their changes. In order to automate the detection of such changes in TSN geospatial files, Theseus embeds an implementation of the TSN Semantic Matching Algorithm that computes LOD semantic graphs describing all the TSN elements and their evolution, based on the vocabulary of the two ontologies.This framework is intended first for the Statistical Agencies, since it considerably helps in complying with Open Data directives, by automating the publication of Open Data representation of their geographic areas that change over time. Second, the created LOD graphs enhance the understanding of territorial dynamics over time, providing policy-makers, researchers, general public with semantic descriptions of territorial changes to conduct various analyses upon their jurisdiction, in time and space. The applicability and genericity of our approach is illustrated by three tests of Theseus, each of them being led on three official TSNs: The European Nomenclature of Territorial Units for Statistics (NUTS) (versions 1999, 2003, 2006, and 2010) from the European Eurostat Statistical Institute; The Switzerland Administrative Units (SAU), from The Swiss Federal Statistical Office, that describes the cantons, districts and municipalities of Switzerland in 2017 and 2018; The Australian Statistical Geography Standard (ASGS), built by the Australian Bureau of Statistics, composed of seven nested divisions of the Australian territory, in versions 2011 and 2016.
Complete list of metadatas

Cited literature [193 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Monday, March 30, 2020 - 11:54:11 AM
Last modification on : Monday, December 7, 2020 - 12:36:03 PM


Version validated by the jury (STAR)


  • HAL Id : tel-02524361, version 1



Camille Bernard. Immersing evolving geographic divisions in the semantic Web. Mathematical Software [cs.MS]. Université Grenoble Alpes, 2019. English. ⟨NNT : 2019GREAM048⟩. ⟨tel-02524361⟩



Record views


Files downloads