Data in Switzerland: BFS at OKCon 2013

Post on 29-May-2015

999 views 2 download

Tags:

Transcript of Data in Switzerland: BFS at OKCon 2013

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Semantic Data!Linked Data Platform@ FSO

Adrian Gschwend (Alain Nadeau)September 2013, OKCon Geneva

2 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Swiss Federal Statistical Office FSO

• Runs and co-ordinates Switzerland’s entire public statistic system

• Produces and disseminates key statistical information in• Demographics, Economy, Society, Spatial planning,

Environment• Complex analysis, interpretation and monitoring of its data

http://www.bfs.admin.ch/bfs/portal/en/index.html

3 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Semantic Data!

• OpenData policy since foundation in 1860, public mandate• More than 50’000 files available as of today• 2 or 3 star format according to Tim Berners-Lee

• CSV, PX, PDF, printed books• FSO follows developments in data dissemination

• Berners-Lee 5-star• Semantic data are the future of open data

• FSO mandates Berner Fachhochschule (BFH) to pilot a study: How can existing statistical data (Census 2011) be transformed into semantic data and linked with other semantic data

http://lod-cloud.net/versions/2011-09-19/lod-cloud_colored.html

4 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

We are not alone

• Wikipedia / DBpedia / Wikidata• Facebook Graph Search• Google Knowledge Graph (Freebase)• JSON-LD & RDFa (microformats)• Linked Data Cloud

Trend: Structured, graph based data!

5 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Linked Data Platform for BFS

• We know our data best• Tabular publication lacks clear semantics• You might interpret a relation differently• Goal: Increase machine readability, clear vocabulary• Standard: Statistical Data and Metadata Exchange (SDMX)

http://sdmx.org/

6 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Technology Stack

• Standardized graph query language: SPARQL• Ask ad-hoc questions without the performance hit

• Output formats: CVS/TSV, JSON, XML & RDF• Direct access via HTTP (data and queries)

7 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

RDF

• W3C standard• Resource Description Framework• Provides various serializations (JSON, XML, Turtle etc.)• Supports internationalization by design (i18n)• Technical base for SPARQL• Federation (no need for one big DB)• In your namespace• Logic languages on top of it (RDFS, OWL)• Support in most popular programming languages

8 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Prototype with FSO data

• Public SPARQL endpoint• Sample apps/visualization• Own namespace (data.admin.ch)

• Unique identifiers for every piece of data • Direct access to HTTP URIs in various

RDF formats (dereferencing)• Sample integration from different data

source (DBpedia, Google Maps)

http://www.gemeindezukunft.ch/acms/projekte/opendata/

9 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Sample Applications

10 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

11 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

HTML view

12 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Turtle Serialization

curl -H "Accept: text/turtle" http://data.admin.ch/bfs/municipality/12099

13 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Turtle output

14 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Same for JSON

Curl -H "Accept: application/rdf+json" http://data.admin.ch/bfs/municipality/12099

15 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

JSON output

16 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

JSON-LD for „classical“ web coders

curl -H "Accept: application/ld-json" http://data.admin.ch/bfs/municipality/12099

Work in progress, not implemented yet on public endpoint!

17 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

JSON-LD output

18 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Ask ad-hoc questions

19 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Get answers your way

20 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Visualize things your way

21 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

22 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

What you get

• The data is the API• Metadata comes with the data

• Standardized vocabulary• Standardized query language• Become your own Data Analyst and Data Scientist• Reduces ETL pain of other formats• Federation (see DBpedia integration)• You choose your output format• Built on HTTP stack (web native!)• Advanced: Schemas, RDFS & OWL

• Ask me for demos

23 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

What you pay for

• Steeper learning curve• Data owner has more work to do upfront• You need to get rid of some old concepts in your head

But it‘s worth it!

24 Semantic data!| OKCon September 2013Gschwend/Nadeau

Federal Department of Home Affairs FDHAFederal Statistical Office FSO

Outlook

• Stable namespace ?• More datasets ?• A productive, redundant platform ?• More government divisions on board?

Talk to us for questions / remarks / participation!