Linked Data at the German National Library

1

Linked Data at the German National Library

Reinhold Heuvelmann

Overview

1. Basis: Data Management at DNB

2. Linked Data Service

- Features

- Infrastructure

- Maintenance

- Data Modeling Workflows

- Usage

- Future

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 2

We work to position the DNB as an important navigation aid in the cultural heritage graph

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 3

Photo

by N

ico K

ais

er

(CC-B

Y):

htt

p:/

/ww

w.f

lickr.

com

/photo

s/n

icokais

er/

4667377944/s

izes/z

/in/p

hoto

str

eam

/

Metadata turntable

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 4

ILS internal format

Pica

IN OUT

Manual cataloguing Machine processing Import: MARC 21 ONIX XMetaDissPlus JATS

MARC 21 ONIX MODS Dublin Core Linked Data BIBFRAME

Traditionally high link density

5

"Work-stance"

Part

"Work-stance" Whole

Work Person

"Workstance": Work+Instance

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016

Linked Data Service - Datasets and profiles -

Authority data Gemeinsame Normdatei

(GND)

Profiles

• GND Ontology

• Entity Facts

Bibliographic data German National

Bibliography

Profiles

• DINI AG KIM

• BIBFRAME (Prototype)

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 6

Established

2010

Linked Data Service - Figures -

– Bibliographic data (DINI-KIM Profile) 218 million triples* Ø 17,5 triples per title

Dump size 15 GB**

– Authority File GND (GND Ontology) 131 million triples* Ø 11,5 triples per entity

Dump size 11 GB**

* as of January 2016 **RDF Serialization Turtle

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 7

… just another export format

– Persistent and dereferencable URIs: http://d-nb.info/ - Identification: http://d-nb.info/1002535506

- Description: http://d-nb.info/1002535506/about/lds

- Content negotiation

– Data dumps

– Retrieval API: SRU

– Harvesting: OAI-PMH

– DNB Data Shop

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 8

Linked Data Service - Unique Features -

– Target audience: non-library users, web context

– Enrichments with links to/from external sources

– 3 Serializations - RDF/XML

- Turtle

- JSON-LD

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 9

Conversion Seriali-

sation

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 10

ILS

Data Conversion Service

Conversion I

n

d

e

x

Enrichment

Database

Wikidata

d-nb.info data dumps SRU OAI DNB Data Shop

VIAF Classification

Alignments

GND-ID-Lists

Linked Data Service - Infrastructure -

What does it mean to be an export format?

– Documentation - Application profiles, Element sets

- Release changes notes

– Continuous development of infrastructure, application profiles

– Channels for communication, user services - [email protected]

- Data service

– Fixed release cycle

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 11

What needs to be done regularly?

– Application Profile maintenance - e.g. with introduction of RDA

- ongoing extension

- towards completeness (bibliographic data)

- provenance

– Links to external resources - identification of new targets (and sources)

- updating from link sources

– Validation / Data quality

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 12

Focus: Data modeling workflows - GND Authority data -

GND Ontology – focus: completeness

– Custom-made element set - New elements introduced as source advance

- For native RDF data usage

- Maintaining alignments to other well established element sets

Entity Facts – focus: keeping it simple - Mainly subsection of GND Ontology

- For direct usage: display, indexing

- Requirement driven

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 13

Focus: Data modeling workflows - Bibliographic data -

Competence center for interoperable metadata (DINI AG KIM)

– Sub working group publishes recommendations on the RDF representation of bibliographic data

– Agreement of German speaking Linked Library Data providers and stakeholders i.e. library networks, national libraries

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 14

DINI AG KIM - involvement

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 15

lobid.org

lod.b3kat.de

lod.hebis.de

dnb.de/lds

ld.zdb-services.de

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 16

(German)

The DINI AG KIM profile

– Flat data model

– Mixture of multiple element sets, applied as assumed most widely understood: - RDF Schema

- Web Ontology Language

- Dublin Core

- Bibliographic Ontology

- RDA unconstrained

- schema.org

- …

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 17

BIBFRAME Profile

– Project status

– Prototype

– focus: gaining experience and supporting/triggering discussion

– update on BIBFRAME 2.0 projected

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 18

Next Steps & Future Development

19

Step 1: More data

Goal: publish the complete national bibliography as linked data:

– Musical scores

– Recordings of music

– …

– At a later stage, we might continue with museum and archive collections

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 20

Step 2: More metadata

Goal: Increase visibility of the datasets

Implement and publish dataset descriptions using DCAT, VoID and ADMS

not only for RDF, but also for MARC 21

to be included in datahub d-nb.info

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 21

Step 3: More links

Goal: Enable cross-domain, cross-language and cross-institution discovery

Publish all links between subject vocabularies

Create and publish links between national bibliographies

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 22

… and beyond ...

– Linked Data API (lookup services?, SPARQL endpoint?)

– RDF @ DNB Portal (schema.org descriptions in HTML)

– Negotiating shapes/profiles over HTTP

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 23

Usage

24

Deutsche Digitale Bibliothek Person and organization pages Integrated Authority File (GND)

Profile Entity Facts, JSON-LD

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 25

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 26

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 27

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 28

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 29

museum-digital Integrated Authority File (GND)

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016

filmportal.de GND Participant: Updates in RDF

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 32

Filmportal

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 33

German National Bibliography: "Empirical Data on Over-length Books"

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 34

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 35

Fun with data!*

*not to be taken too seriously due to data and processing imperfections …

Contact and Support

Product manager Linked Data Service: Jana Hentschke [email protected]

Website: http://www.dnb.de/EN/lds http://www.dnb.de/EN/entityfacts

Mailinglist: [email protected]

| Reinhold Heuvelmann | Linked Data at the German National Library | June 25 2016 36

Thank you!