Hubble Asteroid Hunter III. Physical properties of newly found asteroids
Presentation of the INVENiT Expert Meeting on Monday 16 February 2015
1. Supervision team Academy assistants
Inger Leemans Lora Aroyo Cristina-Iulia Bucur
Department of History Department of Computer Sciences Department of Computer Sciences & Bioinformatics
Chris Dijkshoorn Sebastien Valkenberg Leon Wessels
Department of Computer Sciences University Library Department of History
2. Programme
15.00-15.15 Coffee / Tea
15.15-15.25 Introduction of INVENiT
15.25-15.45 Introduction to search strategies
- The Rijksmuseum interface
- RDF and SPARQL queries
- Using structure in search
- New ways of opening up religious heritage
15.45-16.15 Try-out of the various demonstrators
16.15-17.00 Drinks and evaluation
16. The Rijksmuseum interface
Can we determine how often Noah’s Ark is
depicted in the illustrated bibles of the
Rijksmuseum collection in a certain historical
period?
20. RDF and SPARQL queries
Cristina-Iulia Bucur
Academy assistant
Department of Computer Sciences & Bioinformatics
21. Overview
Resource Description Framework - RDF
State of RDF today
- Semantic Web: vocabularies, ontologies
- Linked Data
SPARQL queries
- Basic rules
- Some examples
22. Resource Description Framework - RDF(1)
- A knowledge representation data model
- Framework for describing resources on the web
- Designed to be read by computers
- Data is in the form of triplets
24. State of RDF today
Semantic web
- Framework to share and reuse data
- Vocabularies, ontologies
- Focused on reasoning
Linked Data
- Method to publish structured data
- Connecting data from different sources
25. Vocabularies
- Define concepts, relationships and constraints
- Easier to read by humans
Examples:
- AAT - Art & Architecture Thesaurus
- ULAN - Union List of Artist Names
- Iconclass
- STCN - Short Title Catalogue of the Netherlands
State of RDF today - Semantic Web
26. State of RDF today - Semantic Web
Ontologies
- More complex and more formal concepts
- Easier to read by computers
Examples:
- RDFS - RDF Schema
- OWL - Web Ontology Language
- SKOS - Simple Knowledge Organization System
- EDM - Europeana Data Model
27. State of RDF today - Linked Data
Large scale
integration
of data
28. SPARQL
Query language for databases
- SPARQL – Sparql Protocol and Rdf Query Language
- Based on the triple model - (subject, predicate, object)
- Query over multiple datasets
29. SPARQL - Basic Rules
PREFIX <- declare schema used
SELECT .... <- values to be returned
FROM ... <- identify data sources
WHERE { ... FILTER ...} <- subject, predicate, object to be
matched against triples of the RDF with or without
constraints
ORDER BY ... <- how the values are to be ordered
31. SPARQL - A first query
Retrieve all prints by Bernard Picart
- Rijksmuseum interface with STCN
Step 1: Find the Bernard Picart we are looking for
Step 2: Retrieve all prints by Bernard Picart
32. Retrieve all prints by Bernard Picart
PREFIX <- declare schema used
SELECT .... <- values to be returned
FROM ... <- identify data sources
WHERE { ... FILTER … } <- subject,
predicate, object to be matched
against triples of the RDF with or
without constraints
ORDER BY ... <- how the values are
to be ordered
33. Retrieve all prints by Bernard Picart
PREFIX <- declare schema used
SELECT .... <- values to be returned
FROM ... <- identify data sources
WHERE { ... FILTER … } <- subject,
predicate, object to be matched
against triples of the RDF with or
without constraints
ORDER BY ... <- how the values are
to be ordered
SELECT DISTINCT ?publicationTitle WHERE {
?publication rdf:type
<http://stcnv.data2semantics.org/resource/vocab/Pu
blication> .
?publication stcn:title ?publicationTitle .
?publication stcn:primary_author ?author .
?publication dcterms:hasPart ?print .
?author stcn:name_full ?authorName .
FILTER regex(str(?authorName), "Picart, Bernard")
}
34. SPARQL - A more complicated query
Show all book titles by Bernard Picart with
their Iconclass themes
Step 1: Query for a person
Step 2: Query for a book
Step 3: Query for a subject matter concept
35. Show all book titles by Bernard Picart with
their Iconclass themes
PREFIX <- declare schema used
SELECT .... <- values to be returned
FROM ... <- identify data sources
WHERE { ... FILTER … } <- subject,
predicate, object to be matched
against triples of the RDF with or
without constraints
ORDER BY ... <- how the values are
to be ordered
36. Show all book titles by Bernard Picart with
their Iconclass themes
PREFIX <- declare schema used
SELECT .... <- values to be returned
FROM ... <- identify data sources
WHERE { ... FILTER … } <- subject,
predicate, object to be matched
against triples of the RDF with or
without constraints
ORDER BY ... <- how the values are
to be ordered
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?authorName ?publicationTitle
?subjectLabel WHERE {
?author rdf:type
<http://stcnv.data2semantics.org/resource/vocab/Author> .
?author stcn:name_full ?authorName .
?publication stcn:primary_author ?author .
?publication stcn:title ?publicationTitle .
?publication dcterms:hasPart ?print .
?print dc:subject ?subject .
?subject skos:prefLabel ?subjectLabel .
FILTER regex(str(?authorName), "Picart, Bernard") .
}
39. Cluster Search
Collection: Rijksmuseum
collection
Ontology: Europeana Data Model
Semantics: Post-query
disambiguation
Audience: Museum website
visitors
Research Space
Collection: British Museum
collection
Ontology: CIDOC-CRM
Semantics: Pre-query
disambiguation
Audience: Researchers
40. Cluster Search
Rijksmuseum Collection
550,000 objects records
160,000 images
Links to thesaury
Graph search algorithm
Demonstrator online
http://sealinc.ops.few.vu.nl/clustersearch/cluster_search
41.
42. Graph Search Algorithm
Match user query with text in database
Find paths to artwork using graph structure
Use path to create clusters
Example query: Rubens
43. Research Space
British Museum Collection
2,000,000 object records
1,100,000 images
Structured query formulation
Relies heavily on CIDOC-CRM ontology
Demonstrator kinda online
http://prototype.researchspace.org
56. Exodus twice
A print bible (illustrations with short narrations of
the original biblical stories): Taferelen der
voornaamste geschiedenissen van het Oude en
Nieuwe Testament, en andere boeken, bij de
Heilige Schrift gevoegt, door de vermaarde
kunstenaars Hoet, Houbraken, en Picart
getekent [...] en met beschrijvingen uitgebreid.
's Gravenhaage, P. de Hondt, 1728 (Translation
of: Figures de la Bible. Amsterdam, B. Picart,
1720.
A Bible with illustrations: Alle de
voornaamste historien des ouden en
nieuwen testaments. Verbeeld in
uytsteekende Konst-platen door den wyd-
beroemden Heer Romeyn de Hooge.
Amsterdam, Jacob Lindenberg, 1703.