sharing data on the web (2013)

Sharing Data on The Web

A Semantic Web Tutorial

James Leigh@jamesrdf

james@3roundstones.comFebruary 2013

★

Towards Linked Open Data

● On the Web with an open license

“It's great to have the data accessible on the Web under an open license, however, the data is locked-up in a document. Other than writing a custom scraper, it's hard to get the data out of the document.”

★★

Towards Linked Open Data

● On the Web with an open license● Machine-readable data

“Splendid! The data is accessible on the Web in a structured way (that is, machine-readable), however, the data is still locked-up in a document. To get the data out of the document you depend on proprietary software.”

★★★

Towards Linked Open Data

● On the Web with an open license● Machine-readable data● Non-proprietary format

“Excellent! The data is not only available via the Web but now everyone can use the data easily. On the other hand, it's still data on the Web and not data in the Web.”

★★★★

Towards Linked Open Data

● On the Web with an open license● Machine-readable data● Non-proprietary format● RDF standards

“Wonderful! Now it's data in the Web. The (most important) data items have a URI and can be shared on the Web. A native way to represent the data is using RDF, however other formats such as Atom can be converted/mapped, if required.”

★★★★★

Towards Linked Open Data

● On the Web with an open license● Machine-readable data● Non-proprietary format● RDF standards● Linked RDF

“Brilliant! Now it's data, in the Web linked to other data. Both the consumer and the publisher benefit from the network effect. ”

Path to Success

● Start easy– Well curated– Small dataset

● Reach out to developers● Get others involved early● Ensure internal benefit● Maximize productivity● Integrate related dataset

Choosing a License

● Allow modification?● Require attribution?● Permit commercial uses?● Must others share alike?● What jurisdiction?

Address Data Quality

● Completeness● Conformity● Consistency● Accuracy● Integrity● Eliminate duplication

★

Put It On The Web

● Upload it and share● Document what is available● Document how to use it

– Solve a consumer need● Encourage feedback

– Continues improvement

★★

Make It Machine-readable

● Simplify data structure● Don't include unstable identifiers● Automate publishing

★★★

Use a Non-proprietary Format

● Standardized formats– CSV instead of excel– XML instead of MDB

● Benefits:– Accessibility– Interoperability– Reduce risk of

● Confidential info● Software viruses

★★★

API Service

● Data access layer● Atom Publishing Protocol● Google Data Protocol● Open Data Protocol● Facebook Graph API● Linked Data

★★★

The Web

● Universal unidirectional links using URLs– Graph of nodes identified by URL

● Cooperation without coordination● It's simple nodes and Links

★★★

The Semantic Web

● A universal scalable interconnection bus● Share and reuse data across

– Community boundaries– Application– Enterprise

★★★

(re)Model the Data

● Identify existing model● Reuse common vocabularies● Model data relationships

– Put aside particular needs of any application● Diagram relationships● Test assumptions with subject matter experts

★★★

Universal Identifiers

● Others can reference things● The Web uses URL for

downloadable things● Uses URI for real world things and

abstract concepts● Two references with the same URI

are the same thing● Quick, easy and scalable look-up

★★★★

Publish RDF

● RDF connects better than any other model● Cooperation without coordination● Data can be combined safely● Mix of vocabularies

★★★★

VoID Description

● De facto standard for describing Linked Data● Used in linked data catalogues● Include provenance and attributions● Associate consumers' license

★★★★

RDF Schema

● Understand structure of the graph– Self-describing

● Define terms and relationships

● Classes of things

★★★★

Simple Knowledge Organization

● Controlled vocabularies● Taxonomies● Thesauri

★★★★

Controlled Vocabularies

● Strongly agree● Agree● Somewhat agree● Neither agree nor

disagree● Somewhat disagree● Disagree● Strongly disagree

● Mr.● Ms.● Mrs.● Miss.● Dr.

● AB● BC● MB● NB● NL● NS● ON● PE● QC● SK● NT● NU● YT

★★★★

Taxonomies

★★★★

Thesauri

★★★★

Web Ontology Language

● Property restrictions● Equivalences across

ontologies● Constraints● Automatic reasoners● Annotations

★★★★★

Linked RDF

● Linked your data to others● Adds more related data● Add value to the data

– By joining it to other data

★★★★★

Generating Data Websites

● Give ordinary users direct access the data● Templates to generate HTML

– Generate static HTML offline– Produce HTML on-demand

Social Responsibility

● Responsibility to maintain published data● Publish frequency of data updates● Have a persistence strategy● Ensure data is accurate as possible● Respond to reports of problematic data

Comments or Questions?