CEDAR census RDF dataset alpha release, v0.1


We’re pleased to announce the first (alpha) release of the Dutch historical censuses dataset as Linked Data. Point your browsers and applications to the SPARQL endpoint


The conversion of the original dataset in Excel has been performed with TabLinker.

The dataset is structured in 507 named RDF graphs, one per census file in the original dataset. Each named graph IRI follows the pattern


Since it would be very verbose to write SPARQL for all these graphs at once, a graph group (warning: Virtuoso exclusive feature and thus non-standard SPARQL) has been created in the triplestore to refer to all these 507 graphs at once, with the IRI


This IRI can be used in the FROM clause of any SPARQL query to launch it against all named graphs of the dataset.

To give some statistics on very preliminary global queries:

Size of the dataset: 110,585,567 triples
Total marked cells: 10,272,862
Hierarchical row header cells: 389,132
Data cells: 7,960,911
Column header cells: 61,110
Row property cells: 3,609
Title cells: 2,150
Row hader cells: 1,581,546
Metadata cells; 274,404

Please note that this release consists of the raw census data only (so no annotations and no harmonization layer yet). Please report all kinds of errors, mistakes or inconsistent results you’ll surely encounter while querying these data.

Big thanks to Michael Schuuring and Jetske van der Schaaf, that worked for several weeks exclusively on the arduous task of marking-up the tables. Big thanks also to DANS for hosting the CEDAR LOD server.