how the web of data will be won

19
How the Web of Data Will be Won John Sheridan Jeni Tennison

Upload: jeni-tennison

Post on 29-Nov-2014

1.910 views

Category:

Technology


2 download

DESCRIPTION

What does it take to create a web of government Linked Data? The UK government is finding out. Our story is one of pioneers. You will hear how we are moving out of existing settlements to the wide plains of government data. How we are starting to build the first railroads across this vast territory to open a new lands of opportunity. All the time, of course, having to avoid both outlaws and the Civil War back east.

TRANSCRIPT

Page 1: How the Web of Data Will be Won

How the Web of Data Will be Won

John SheridanJeni Tennison

Page 2: How the Web of Data Will be Won

Overview

• Mapping Territory

• Laying Tracks

• Gold Mining

• Civil War

• Winning the Web of Data

Page 3: How the Web of Data Will be Won

Mapping Territory

photo from Cornell University Library on flikr

Page 4: How the Web of Data Will be Won

Open Gov't Data

• Pioneers

• Wide open plains

• data.gov.uk

• Our legacy?

Page 5: How the Web of Data Will be Won

Why Linked Data

• "web data"

• Publishers and consumers

• Open standards

• Distributed data

• Small pieces loosely joined

Page 6: How the Web of Data Will be Won

Our Approach

• Winchester '73

• Design patterns

• Try and evolve

• Learning from mistakes

Page 7: How the Web of Data Will be Won

Laying Tracks

photo from Cornell University Library on flikr

Page 8: How the Web of Data Will be Won

URIs

• Things, documents, definitions, datasets

• Recommendations for persistence

• Initial URI sets: legislation, schools, geographies ...

http://{sector}.data.gov.uk/id/{concept}/{id}http://{sector}.data.gov.uk/doc/{concept}/{id}http://{sector}.data.gov.uk/def/{scheme}/{concept}http://{sector}.data.gov.uk/data/{package}/{subset}

Page 9: How the Web of Data Will be Won

Versioning

• Multiple sources, multiple versions over time

• Named graphs and metadata

• dates and relations to other versions

• authority

• source and provenance

• Time-based slices of data

Page 10: How the Web of Data Will be Won

Provenance

• Reproduceability as the basis of trust

• Hugely complex

• origination

• processing

• validation

• Applies to real-world artifacts as well as data

Page 11: How the Web of Data Will be Won

Gold Mining

photo from http://www.archives.gov/research/american-west/

Page 12: How the Web of Data Will be Won

Statistics

• Rich seam of data

• SDMX from eg Office for National Statistics

• Excel spreadsheets

• Pattern for publishing statistics in RDF

• Tools to create linked data from Excel

• http://groups.google.com/group/publishing-statistical-data

Page 13: How the Web of Data Will be Won

Geo-spatial Data

• Tie in with INSPIRE European Directive

• spatial objects must have identifiers (URIs)

• specific metadata about spatial objects

• Publication of geometries (eg boundaries)

•http://www.terrafuture.com/

Page 14: How the Web of Data Will be Won

Civil War

photo from ♪_Lisa_♪ on flikr

Page 15: How the Web of Data Will be Won

Linked Data API

• Neglect usability at our peril

• ease of querying

• ease of processing

• Layer processing on SPARQL endpoint

• create developer-friendly APIs

• More later this afternoon...

Page 16: How the Web of Data Will be Won

Other Services

• Resolution

• searching for the right URI

• Enrichment

• marking up text with UK Government terms

• Backlinking

• Finding pointers from the rest of the cloud

Page 17: How the Web of Data Will be Won

Winning the WoD

photo from http://www.archives.gov/research/american-west/

Page 18: How the Web of Data Will be Won

Winning the WoD

• For everyone

• Brutally practical

• Doing "stuff" matters

Page 19: How the Web of Data Will be Won

Conclusions

• Early days

• Making progress

• Come join us