The CEDAR Project: Harmonizing the Dutch Historical censuses in the Semantic Web

TY - CONF

T1 - The CEDAR Project: Harmonizing the Dutch Historical censuses in the Semantic Web

AU - Ashkpour, Ashkan

AU - Meroño-Peñuela, Albert

PY - 2014

Y1 - 2014

N2 - Historical censuses are on the most consulted, reliable and large scale statistical data sources available which give an insight in the population characteristics of a nation: they provide a wealth of data on many issues in the course of time at the demographic, social and economic level. In the Netherlands, the digitization of the Dutch historical censuses (1795-1971) has resulted in many dispersed tables, and meaningful historical information is currently hidden in aggregated data over 2,300 of these tables. In order to fully reap the benefits of this dataset temporal comparisons are required. However, the Dutch historical censuses are still very difficult to compare, aggregate and query in a uniform fashion due to the lack of data harmonization. The CEDAR project of the Computational Humanities Programme aims at enabling greater access and use of this dataset by applying a specific datamodel (exploiting the Resource Description Framework RDF technology), to create Linked Census Data and make it interlinkable with other hubs of historical socioeconomic and demographic information; and various harmonization practices. The process of harmonization is currently an ongoing process and will be so in the future. In order to implement harmonization on this dataset we have applied a three tier model, distinguishing between the raw data layer (containing a direct translation of the numbers and concepts in the tables), the annotations layer (with corrections and comments from the dataset curators), and the harmonization layer (linking and standardizing heterogeneous census entities). Making extensive use of semantic technologies and Linked Data principles, we make this Linked Census Data comparable over time by leveraging specific data classifications, taxonomies and ontologies. By querying these, we create visualizations in order to explore the thousands of variables and to get an early understanding of our data and its relations. We create bottom up classifications for information contained in all three census types (population, houses and occupations) like demographical structures, housing types, occupational classes and statuses, religious denominations, and so on. We use animation techniques to display the conceptual changes that modified the social landscape in fundamental centuries of Europe's history.

AB - Historical censuses are on the most consulted, reliable and large scale statistical data sources available which give an insight in the population characteristics of a nation: they provide a wealth of data on many issues in the course of time at the demographic, social and economic level. In the Netherlands, the digitization of the Dutch historical censuses (1795-1971) has resulted in many dispersed tables, and meaningful historical information is currently hidden in aggregated data over 2,300 of these tables. In order to fully reap the benefits of this dataset temporal comparisons are required. However, the Dutch historical censuses are still very difficult to compare, aggregate and query in a uniform fashion due to the lack of data harmonization. The CEDAR project of the Computational Humanities Programme aims at enabling greater access and use of this dataset by applying a specific datamodel (exploiting the Resource Description Framework RDF technology), to create Linked Census Data and make it interlinkable with other hubs of historical socioeconomic and demographic information; and various harmonization practices. The process of harmonization is currently an ongoing process and will be so in the future. In order to implement harmonization on this dataset we have applied a three tier model, distinguishing between the raw data layer (containing a direct translation of the numbers and concepts in the tables), the annotations layer (with corrections and comments from the dataset curators), and the harmonization layer (linking and standardizing heterogeneous census entities). Making extensive use of semantic technologies and Linked Data principles, we make this Linked Census Data comparable over time by leveraging specific data classifications, taxonomies and ontologies. By querying these, we create visualizations in order to explore the thousands of variables and to get an early understanding of our data and its relations. We create bottom up classifications for information contained in all three census types (population, houses and occupations) like demographical structures, housing types, occupational classes and statuses, religious denominations, and so on. We use animation techniques to display the conceptual changes that modified the social landscape in fundamental centuries of Europe's history.

KW - census data

KW - linked data

KW - dutch history

M3 - Abstract

T2 - Digital Humanities Benelux Conference 2014

Y2 - 12 June 2014 through 13 June 2014

ER -