What is Semantic Web

• Humans can understand web content
• Machines (computers) need some help
• In a semantic environment Machines can understand the meaning of a thing (document, port, text, system status etc) and link to/from it.
• Metadata are a type of semantics

What do Humans read

            Drupal meetup Thessaloniki will take place
            in Techministry
            at Wednesday 16 March 2016
            from 19:00 to 21:00.

          

What do Machines read

            Word word link word word word
            word link
            word link
            word link.

          

What should Machines read

    -> Event description with url http://mydrupal.gr/node/2881

    Word word Link of a City word word word
    word Link of a Place
    word link of a Date
    word link of a 24h Time.

          

What is Linked Data

• Really connected structured web
• Bizer, Heath and Berners-Lee (2009)
• Internet of Things not only of documents
• Linked Data may not be Open Data
• Automatic decisions
• Find information faster, easier and with accuracy
• The basic 'engine' behind Big Data
• The key for Open Government/Cities/Health/...

The 5th star of the Open Data

Linked data simplified

One giant global graph

RDF

• A graph data model for describing things
• Describe the relationships between things
• W3C specification (1997)
• Only describes resources
• It's a data model, a concept not a data format
• It needs serialization
• Current version RDF 1.1 (2014) after 1.0 (2004)

The RDF triples

• [subject] [predicate] [object]
• "John knows Mary"
• We must state explicitly the nature of the connection
• We can download or query triples from triplestores
• Types of RDF triples: Literal Triples and RDF Links

The RDF Anatomy

• [subject]: URI, Blank_node (bnode)
• [predicate]: URI
• [object]: URI, Blank_node, Literal

RDF - URI

      Absolute IRI which may include a # fragment.
      <http://www.example.org/>
      <http://www.example.org/#fragment>

      Relative IRI resolved against base IRI.
      <abc.rdf>

      Base IRI, usually the query document IRI
      <>

      IRI shorthand using XML-style prefix ex and local name.
      Declared with PREFIX (SPARQL) or @prefix (Turtle)
      ex:name
    

RDF - Literal

      A Unicode string with an optional language tag.
      "hello"
      "bonjour"@fr
      "1234"
          

RDF - Typed Literal

      A Unicode string and datatype IRI for encoding datatypes.
      "1234"^^<https://www.w3.org/2001/XMLSchema#string>

      Abbreviated with an XML QName style as:
      "1234"^^xsd:string

      Short forms for several common datatypes:
      -10

      "-10"^^xsd:integer

      1.2345
      "1.2345"^^xsd:decimal

      true
      "true"^^xsd:boolean
          

Taxonomies, Vocabularies and Ontologies

• Domain-specific terms for describing classes (groups) of things and how they relate to each other
• Lightweight ontologies in RDF often referred as vocabularies
• They borrow classes/properties from each other
• We can create our own, extend or reuse ontologies
• Popular: Schema.org, SKOS, FoaF, DCMI, SIOC etc
• Overview of Linked Open Vocabularies at LOV

RDF - Namespaces and Prefixes

Find prefixes at prefix.cc

The RDF Graph

• A collections of statements about a thing
• Starting with the same Subject
• The URIs occurring as subject and object are the nodes in the graph
• A real example of a webpage

Serializing RDF data

• RDF/XML (.rdf)
• RDFa
• Turtle (.ttl)
• N3, Notation 3 (.n3)
• N-Triples (.nt)
• RDF/JSON (.jsonld)

Example RDFa: demos, real webpage, webpage parsed

SPARQL 1.1 (2013)

• 'Default' query language for RDF (there are many others)
• Similar to SQL
• Fast and secure
• Endpoints, frontends, wrappers
• Apache Jena, Virtuoso, Marmotta, Dydra and many others
• Can Combine several endpoints (federated query)
• Can Update RDF Graph stores (Update query)

SPARQL - Reference synopsis

SPARQL - Common Syntax

    # prefix declarations
    PREFIX foo: <http://example.com/resources/>
    ...
    # dataset definition
    FROM ...
    # result clause
    SELECT ... ?variables
    # query pattern
    WHERE {
        ... ?variables
    }
    # query modifiers
    ORDER BY ...
    LIMIT n OFFSET m
          

SPARQL - Examples

• rdf.myexperiment.org (demos)
• query.wikidata.org (real)
• open-data.europa.eu/en/linked-data (real)
• s3space.com (real)

SPARQL - Tips

• Same results can be achived with different queries
• More accurate queries are faster
• Prefer Group Graph patterns instead of FILTER
• Avoid SELECT *
• Avoid ORDER BY
• Avoid DISTINCT (use REDUCED)
• Use OFFSET, LIMIT to paginate results
• Variables are case insensitive (?var is the same as ?VAR)
• Variables cannot change name on the same query
• Variables can start with ? or $

SPARQL - Cheatsheets

• Semantic Web Reference Card v2.0
• SPARQL RDF Query Language Reference v1.8
• SPARQL PROTOCOL AND RDF QUERY LANGUAGE