gesisDataSearchKG- Semantic search over Social Sciences Dataset

Description

gesiDatasSearchKG is an initiative to extend the current Gesis Data Search to semantic search over KG. This version models the sample of 11,965 datasets available in Gesis Data Search represented in the DDI standard and lifts them to the RDF/S model. The comprehensive data model comprises dataset properties ubiquitous to all domains and specific to social sciences to mitigate the challenges of integrating different DDI versions due to varying elements. We exploit established vocabularies such as schema.org, DISCO, PROV-O, and DCAT, adhering to the best practice of ontology reuse.

gesisDataSearchKG consists of a multilingual dataset belonging to 11 study groups with publication dates ranging from 1966 to 2023. We model 33 semantic properties of the Dataset with 530k triplets.

Data model

In the following figure, the data model of gesisDataSearchKG is illustrated.

Vocabularies

Name	Vocabulary and Type	Domain	Range
StudyGroup	disco:StudyGroup, prov-o:Entity	-	-
DataCatalog	schema:DataCatalog, prov-o:Entity	-	-
CatalogRecord	dcat:Class:Catalog_Record, prov-o:Entity	-	-
Dataset	schema:Dataset, prov-o:Entity	-	-
Organization	schema:Organization	-	-
dataset	schema:dataset	DataCatalog	Dataset
record	dcat:record	DataCatalog	CatalogRecord
wasGeneratedBy	prov-o:wasGeneratedBy	CatalogRecord	CatalogRecordOrganization
inGroup	disco:inGroup	Dataset	StudyGroup
name	schema:name	StudyGroup, CatalogRecordOrganization, Organization, Dataset, Place, Language, Grant_Organization	Text
identifier	schema:identifier	Dataset	Text
alternateName	schema:alternateName	Dataset, Language	Text
creator	schema:creator	Dataset	Organization, Person
publisher	schema:publisher	Dataset	Organization, Person
contributor	schema:contributor	Dataset	Organization, Person
abstract	schema:datePublished	Dataset	Text
citation	schema:citation	Dataset	Text
copyrightNotice	schema:copyrightNotice	Dataset	Text
temporalCoverage	schema:temporalCoverage	Dataset	Text
about	schema:about	Dataset	Text
spatialCoverage	schema:spatialCoverage	Dataset	Place
collectionMode	disco:collectionMode	Dataset	CollectionMode
endDate	schema:endDate	CollectionMode	Text
startDate	schema:startDate	CollectionMode	Text
description	schema:description	CollectionMode	Text
funding	schema:funding	Dataset	Grant
funder	schema:funder	Grant	Grant_Organization
version	schema:version	Dataset	Text
inLanguage	schema:inLanguage	Dataset	Language
conditionsOfAccess	schema:conditionsOfAccess	Dataset	Text
comment	schema:comment	Dataset	Text
sameAs	prov-o:sameAs	Place	URI
hasEmbargoDuration	disco:hasEmbargoDuration	Dataset	Text
analysisUnit	disco:analysisUnit	Dataset	Text
kindOfData	disco:kindOfData	Dataset	Text

Availability of Features

All properties associated with the social science study and their distribution. Availability in Percentage denotes the percentage of total datasets (11965) that these properties are included.
Properties	Availability in Percentage	Description
study group	100	Study grouping information.
Catalogrecord creator	100	Metadata provider details.
version	100	Version number.
title	100	Study title.
copyright notice	78.7	Copyright text.
published date	100	Publication date.
language	100	Study language.
research identifier	100	DOI.
abstract	85.34	Study summary.
comment	100	Comments section.
citation	50.35	Study references.
creator	46.62	Study creator.
publisher	99.9	Publisher details.
contributor	63.01	Contributor details.
topical coverage	68.3	Study topics.
spatial coverage	70.6	Study locations.
temporal coverage	68.3	Time period.
collection mode description	15.5	Data collection methods.
collection start date	100	Start date of data collection.
collection end date	66.21	End date of data collection.
funding information	0.008	Funding details.
analysis unit	17.18	Analysis scope.
universe	60.12	Target population.
kind of data	100	Data type (text, recording, etc.).
sub title	5.53	Alternate title.
embargo date	0.36	Embargo details.

Knowledge Graph Overview

Statistic	Value
Total Number of Triples	527819
Total Number of Unique Entity Types	9

Dataset

To cite this resource, please use the following DOI: 10.5281/zenodo.11070841.

SPARQL endpoint

A SPARQL endpoint is available to send SPARQL queries and retrieve results from gesisDataSearchKG.

https://data.gesis.org/gesisdatasearchkg/sparql

SPARQL Queries

Example 1: Requesting number of studies published per year. ( Result)

					
						SELECT  DISTINCT   (year(xsd:dateTime(?date)) AS ?year) (COUNT(?survey) AS ?numberOfStudies)
						WHERE {
								?survey a .

							?survey  ?date.
									

							}
							GROUP BY YEAR(xsd:dateTime(?date))
							ORDER BY YEAR(xsd:dateTime(?date))

Example 2: Requesting number of studies belonging to 'ALLBUS' per year. ( Result)

					 
						SELECT  DISTINCT   ?date (COUNT(?survey) AS ?numberOfStudies)
						WHERE {
							 ?survey a .
							?survey  ?Group.
							?Group  ?group.
							?survey  ?subject.
							?survey  ?date.
							filter( regex(?group, "ALLBUS" ))
						
						}
						GROUP BY  ?date
						order BY ?date

Example 3: Requesting studies related to national identity in Germany, performed individually . ( Result)

						 
							SELECT ?survey ?title ?Date ?country ?subject ?analysisunit
  
							WHERE { 
							
						  ?survey a .
						  
						  ?survey  ?title .
						  
						  ?survey  ?subject.
						  
						  ?survey  ?country .
						  
						  ?survey  ?Date.
						  
						  ?survey  ?analysisunit.
						  
						  FILTER(    
						  
							REGEX(LCASE(STR(?subject)), "national identity|nationale identität", "i") && 
							
							REGEX(LCASE(STR(?analysisunit)), "individual|individuell", "i") 
							
						  ) 
						}

URLs for Organization names evaluated can be download via Link.

Dump of ntriples Link.

gesisDataSearchKG