Sw 2

• Semantic Web Lecture C02 focuses on Linked Data Basics
• The course aims to teach how to identify, describe, and access descriptions of data

Creative Commons Licensing

• The slide set is part of the Semantic Web Technologies lecture at RWTH Aachen University
• Lars Gleim prepared these slides which are based on the work of Benedikt Köhler and Andreas Harth
• The material is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0): http://creativecommons.org/licenses/by/4.0/

Foundational Literature

• Foundations of Semantic Web Technologies by Hitzler, Krötzsch, and Rudolph (2009) provides semantic web foundations
• Knowledge Graphs by Hogan et al (2021)
• Linked Data by Heath and Bizer (2011) explores semantic web theory and technology

Introduction to Linked Data Recap

• You should already be familiar with how the internet and web function
• URIs are used to identify resources on the web
• HTTP enables access to resource representations in various formats
• Information resources can be identified and these resources represent files

Lecture Agenda

• Semantics on the Web
• Linked Data Principles
• Resource Description Framework (RDF)
• Turtle Syntax for RDF
• JSON-LD Syntax for RDF
• Connecting Things with Documents

Data Models Overview

• Data models formalize objects within a database and their relationships, forming the database schema
• Key aspects covered include structures (object types and their relationships), operations (extraction, linking and query languages such as SQL, OQL, XQuery and SPARQL), and constraints (model inherents and user-defined restrictions)

Data Model Types

Structured

• Defined by rigid table structures with rows (tuples) and columns (attributes), or by objects with attributes and methods

Semi-structured

• XML uses structural elements (tags) that include unstructured content
• RDF uses labeled nodes and edges, with URIs or literals as labels

Unstructured

• BLOBs: possess unrecognizable internal structure
• Key-Value Stores (NoSQL): use arbitrarily composed data elements

Relational Data Model

• It is based on sets (tables, relations)
• Suggested by Edgar F. Codd in 1970
• IBM System R in 1980 was the first commercial implementation

Semantic Web Idea

• Tim Berners-Lee, James Hendler, and Ora Lassila published "The Semantic Web" in Scientific American (2001)
• The article envisions the evolution of the internet to the Web of Data

Wikipedia Web

• Graphically visualizing Wikipedia pages and hyperlinks reveals interconnectedness
• Each dot symbolizes an article; thematic color-coding is employed
• It is possible to locate where the city of Aachen is located within this web
• The web has a high amount of inlinks and outlinks
• The question becomes how to meaningfully utilize this information

Accessing Information

• Consider the question "What is Aachen's largest district by population?"
• Humans intuitively break this down into smaller questions
• This contrasts with web's use of string maching

Limitations of String Matching

• String matching is a fast retrieval method that can get messy

Information Representation in Text

• Aachen is described as a city in North Rhine-Westphalia, Germany with a population of 247,380

Representing Information as Graphs

• Graph representation structures information using subject-predicate-object triples
• Aachen examples: "Aachen is a big City", "Aachen is located in North Rhine-Westphalia", etc

Querying Triples

• Data represented as triples are queriable
• SPARQL, the query language, will be talked about later

Web Standards

• RDF provides a graph-based data model utilizing nodes and arcs
• RDF identifies objects using URIs and represents interlinked information (relationships)
• Vocabularies or Ontologies provide shared understanding, organize knowledge for machine comprehension, and give data an exploitable meaning

Semantic Web "Layer Cake"

• The layer cake consists of: User Interface & Applications, Trust, Proof, Unifying Logic, Ontology (OWL), Rule (RIF), Query (SPARQL), RDF, XML and URI/IRI

Why Graphs?

• Graphs effectively represent relationships: (Oche knows Aachen) and properties (Aachen has a label)

Life Sciences Application

• Life sciences use graphs to represent relationships and interactions within gene and pathway databases

Linked Data Principles Overview

• Semantics on the Web
• The Linked Data Principles
• The Resource Description Framework (RDF)
• Turtle Syntax for RDF
• JSON-LD Syntax for RDF
• Connecting Things with Documents

Linked Data Principles Postulate

• Postulated by Tim Berners-Lee in 2006 stating that the semantic web is about linking data so that machines can explore related data

Tim Berners-Lee's Linked Data Principles

• Use URIs as names for things
• Use HTTP URIs so that people can look up those names
• Provide useful information using standards (RDF, SPARQL) when a URI is looked up
• Include links to other URIs to enable discovery of related things

Linking Open Data Cloud

• The LOD cloud diagram contains hundreds of datasets across many domains
• It provides a good starting point to understand available data

DBpedia Highlights

• DBpedia contains data extracted from Wikipedia infoboxes for structured knowledge
• The resulting RDF dataset is available as Linked Data
• This project is community-led and both the extractions scripts and the data are availabe under open licenses
• These data also links to many other Linked Data sites

DBpedia Example

• DBpedia extracts data from Wikipedia articles, as exemplified by Neil Young's data

Wikidata

• Wikidata is a Wikimedia Foundation project started by the creators of Semantic MediaWiki
• It aims to provide a knowledge base for generating fact boxes across different Wikipedia language versions
• All data is openly licensed

Principle 1: URIs as Names

• Point to a distinct resource when sharing information
• Linked Data follows a resource-centered view of data modelling
• Resources form the basis of web architecture
• A book could be identified by its ISBN using the RFC 3187 scheme. For example, urn:isbn:9-781497-364783.

Principle 2: Use HTTP URIs for Lookup

• Utilize HTTP as a mechanism to retrieve more information about a thing, given its URI
• Requires a mapping between a URI as a name (identifying an object) and a URI as a location (identifying a machine-readable description

Principle 2: ISBN as Example

• Information can be found through the following ways:
  • Lookup in a local bookstore by ISBN
  • Lookup in an online bookstore by ISBN
  • Query an ISBN identified database

Principle 2: Inherent Lookup

• HTTP URIs inherently unite logical and physical addresses which allows for immediate information retrieval
• Using a website of Tim Berners-Lee as example you can view the page content immediately
• Accessing information requires no mediator
• Content is serialised into bytes

Principle 3: Information Provision

• Provide useful information with standards whenever someone looks up a URI
• Content becomes meaningful to computational users and enables Linked Data
• RDF (Resource Description Framework) facilitates content's utilization
• Serializations of RDF include RDF/XML (widely used, tedious to read and write), Turtle, Notation 3, and JSON-LD

Principle 4: Linking

• Associating items from different sources creates a mesh for performing algorithms
• Links connect separate data graphs
• A graph-structured data model and URI re-use allows easy merging
• DBpedia is an example fo URI usage with fast merging

RDF and Turtle Overview

• Semantics on the Web
• The Linked Data Principles
• The Resource Description Framework (RDF)
• Turtle Syntax for RDF
• JSON-LD Syntax for RDF
• Connecting Things with Documents

RDF

• Framework used for resource description
• A W3C Recommendation since 1998
• Data model used for describing metadata of web resources
• Encodes structured information in a universal format in graphs
• Described using Vertives and edges

Key RDF Concepts

• RDF is foundational for both Semantic Web and Linked Data
• An RDF triple describes information with a subject-predicate-object structure
• Predicates specify the relationship in subject and object
• Triples viewed graphically

RDF Terms

• RDF terms consist of URIs, blank nodes, and literals which are disjointed which can be used to create triples
• URIs globally identify resources
• Blank nodes identify resources only within a file
• Literals refer to data values

RDF Graphical Representation

• RDF triples can hold different types of elements, which are represented differently

RDF Graph Example

• RDF triples, form directed labeled graphs

RDF/XML Example

• An example of what the RDF/XML markup code looks like

Turtle Overview

• Semantics on the Web
• Linked Data Principles
• Resource Description Framework (RDF)
• Turtle Syntax for RDF
• JSON-LD Syntax for RDF
• Connecting Things with Documents

Turtle Syntax

• Is a terse RDF triple language which is text based, and is recognized by the W3C
• It is also a subset of Notation 3 and RDF syntax

Turtle and Triples

• Offers a simple means is writing triples
• Subject, predicate and object structure is used with a dot after
• URIs identify resources and enclosed in angle brackets
• The subject predicate object representation becomes simple

CURIEs

• CURIEs (Compact URIs) offer a way to abbreviate URIs
• The keyword @prefix is used
• With the correct prefix, CURIEs can be expanded

Turtle with CURIEs

• Prefixes has to be defined in each file
• CURIEs composed of prefix and reference pairs and written with prefix and no brackets

CURIEs and Links

• Last slides triples refer to resources in each document
• URIs are invalid when document moves
• Relative default prefixes points to document's location

URIs and Links

• HTTP URIs occurs in relative forms
• References to URI must be hierarchical
• Can use notations for file systems

Typed Literals

• Its necessary to link information
• Literals can be types with XML Schemas

Data Types

• Literals are untyped as strings and typing allows for semantic interpretation of values
• URIs are chosen for ID and there is syntax
• rdf:HTML and rdf: XMLLiteral are predefines

Datatypes

• Description and Example
• xsd:string Character string Tim
• xsd:boolean Boolean true
• xsd:integer Arbitrary-side integer numbers 5
• xsd:int 32-bit integer numbers 5
• xsd:decimal Real numbers which can be represented by decimal numbers 3.14
• xsd:float IEEE single-precision 32-bit floating point 3.14
• xsd:double IEEE double-precision 64-bit floating point 3.14
• xsd:dateTime Date and time inspired by ISO 8601 2016-06-18T23:59:59.123+00:00

Specifying Literals

• To specify a literal's datatype use ^^
• Turtle uses abbreviations

Typed Literals Examples

• Showing population and average temperature: includes prefix

Repeating Subject

• Can add more using the terms terse"
• Using predicate or object relates can go in using semicolons

Repeating Subject and Predicate

• In a case where a lot of objects correlate, use the comma to denote relation

Turtle w/ Shortcuts

• Finish what was said using single dot and shorter codes

Language Tags

• Characters denote literal's language in text by the @
• Languages have tags for the literal

Blank Nodes

• Simply state existence
• Labels are at beginning
• Triple shows resource and code

Using Notes

• Several can be coded
• Stating some one has "Bob" in their label
• Has to be in the object and has to have an external source

Nesting Nods

• In use or reuse, Turtle allows for nesting
• With the layout, a fresh layout uses surrounding or correlating pairs

N - ary Relations

• Indicate relationship of more than two resources
• Are like binary, using a lot of notes
• Often identifies with the tool as a resource"

Example

• Has specific information about prefix and the year, to show the years population with correct coding

Links and Containers

• Common data and enumeration, use many resources: lists the different types of data and collections

Container Types

• Indicate the rdf:types.
• rdf:Seq - Used as ordered list, sequence
• rdf:Bag - Used as unordered set
• rdf:Alt - Used as options and list of element

Example Links

• This example codes lists their name and their specific ID/Name to relate back

Turtle and Syntax

• Short cuts in code helps to express lists
• Object is Rdf collection blank node
• This short cut helps with creating RDF value for this part of the code