Practical Cross-Dataset Queries with SPARQL (Introduction)

Post on 14-Jan-2015

2.132 views 2 download

description

Introduction slide deck for the tutorial on “Practical Cross-Dataset Queries for the Web of Data” we presented at WWW2012. The tutorial homepage, including other presentations, is here: http://latc-project.eu/events/www2012-tutorial-cross-dataset-queries

Transcript of Practical Cross-Dataset Queries with SPARQL (Introduction)

Practical Cross-Dataset Queries on the Web of Data

Tutorial @ WWW2012, Lyon, FranceRichard Cyganiak, Knud Möller, Anja

Jentzsch, Andreas Schultz, Robert Isele, Pablo Mendes

The Web is becoming a platform for data exchange.

• Microdata, Schema.org, web APIs, Linked Data Cloud, Open Data movement, …

• Often need to combine local and remote data from several heterogeneous sources

• Scripting and mash-ups. This works, but can we do better?

SPARQL as a query languagefor the Web

• Data from all of these data sources can be converted to RDF using off-the-shelf tools, or the sources are already RDF.

• SPARQL is W3C's standard query language for RDF

• SPARQL 1.1 just out, great new features for working with heterogeneous data

Caveats

• We will focus on ad-hoc queries.• This is not just about what works, but also

about what doesn't work.

How to get data into RDF format

• Relational: R2RML standard; D2RQ, Virtuoso RDF Views, Revelytix Spyder

• Excel, CSV: RDF Extension for Google Refine, XLWrap

• XML: XSPARQL• JSON: JSON-LD• Microformats, Microdata: Apache Any23• Collect data from many web pages: LDSpider

SPARQL: The big picture

SPARQL client

SPARQL engine

RDFStore

SPARQL Protocol

Scenario: Remote SPARQL endpoint

SPARQL client SPARQL engine

RDFStore

Scenario: Local SPARQL store

SPARQL client

SPARQL engine

Scenario: Local SPARQL engine, load data from files on the fly, no store

LocalRDFfile

RemoteRDFfile

Non-RDFfile

Conversion

SPARQL engine

RDFStore

Scenario: CONSTRUCT the input data

SPARQL client

SPARQL engineLocalRDFfile

SPARQLCONSTRUCT

query

SPARQL engine

RDFStore

LocalRDFfile

SPARQLCONSTRUCT

query

SPARQL engine

RDFStore

Scenario: Federated Query

SPARQL client

SPARQL engine

Basic Federated Query

LocalRDFfile

… or any combination of these.

Agenda – Morning

• Linked Data Basics• SPARQL Basics• 10:30–11:00 Coffee• Federated queries with SPARQL• Hands-on session 1• 12:30–13:30 Lunch

Agenda – Afternoon

• 12:30–13:30 Lunch• Schema mapping with SPARQL CONSTRUCT• Instance matching with Silk• Finding RDF datasets• 15:00–15:30 Coffee• Visualizing SPARQL query results• Hands-on session 2• 17:00 Adjourn

Hands-on sessions

• USB sticks with data, queries, and instructions• Install Apache Jena command line tools• Need a browser with a JavaScript console

(recommended: Firefox+Firebug or Chrome)

Music

Presenters

• Richard Cyganiak, DERI• Knud Möller, Talis• Anja Jentzsch, FU Berlin• Andreas Schultz, FU Berlin• Robert Isele, FU Berlin• Pablo Mendes, FU Berlin• (Christophe Guéret, VUA)• (Michael Hausenblas, DERI)

Please interrupt and ask questions!