The document discusses the integration and exploration of financial data using semantic web technologies, focusing on mapping XBRL to RDF to improve data comparability. It highlights the development of semantic tools and ontologies for financial data integration, illustrating a project that generated over 125 million triples from XBRL filings. Despite initial promise, the project was discontinued in 2013 due to concerns about the effectiveness of the applied semantics.

1. Integration and Exploration of Financial
Data using Semantics and Ontologies
Roberto García
HCI and Data Integration Research Group
Universitat de Lleida, Spain
8
Academic Track

2. Introduction
• Report experience first attempt to map “significant”
amount of XBRL to RDF
– 2008 +100M triples… and as far as we know
• Test Semantic Web and Linked Data for financial
data integration and exploration
• Convert it, map XBRL to RDF
– Apply generic mapping: ReDeFer
• XSD2OWL, XML2RDF
• Once as RDF+OWL, improve comparability using
semantic links

3. Semantic XBRL
• Dataset size (from US SEC)
– May 2012: 125,55 million triples from 29342 XBRL filings
– https://datahub.io/dataset/semantic-xbrl
– Linked Open Data Cloud
0
20
40
60
80
100
120
140
0
5000
10000
15000
20000
25000
30000
35000
Filings
Triples
(millions)
Publishing XBRL as Linked Open Data
LDOW’09 @ WWW’09
Position Paper @ 2009 W3C Workshop on
Improving Access to Financial Data on the Web
Linking XBRL Financial Data (chapter)
Linking Enterprise Data Book, 2010
Using Semantic Web Technologies to Facilitate
XBRL-based Financial Data Comparability
FEOSW'12 @ ESWC’12
Improving the comparability of financial statements in
a changeable context: a Semantic Web-based approach
PhD Thesis, Economics. Héctor Carretié, 2012

4. Approach
XML2RDF
XSD2OWL
XBRL Schemas
.xml
XBRL filings
XBRL Ontologies
ReDeFer
Rhizomer
+1M triples
LOD
Linking
RDF2Form
XHTML+
RDFa
RDF2HTML
SPARQL
edit
post
Rhizomer
AJAX
get
.xsd
EDGAR
RSS Feed
download
EDGAR
Standard
Taxonomies
.US
+100M triples

5. XBRL XML to RDF
• ReDeFer XML2RDF,
model XML tree using triples
– xsd:element and xsd:atribute  rdf:Property
– xbrli:id and xbrli:identifier  rdf:Resource ID
– Other resources, anonymous
Rendered by RDF2SVG

6. XBRL XSD to OWL
• XBRL Schemas: XBRL 2.1, US GAAP1,
Spanish PGC…
• ReDeFer XSD2OWL
XML Schema OWL
element | attribute
rdf:Property
owl:DatatypeProperty
owl:ObjectProperty
element@substitutionGroup rdfs:subPropertyOf
element@type rdfs:range
complexType owl:Class
complexType//element owl:Restriction
extension@base |
restriction@base
rdfs:subClassOf
@maxOccurs,
@minOccurs
owl:maxCardinality,
owl:minCardinality
sequence,
choice
owl:intersectionOf,
owl:unionOf
Core classes and properties for XBRL Instance
XSD2OWL mappings
1 Ontologies available from http://rhizomik.net/ontologies/bizontos

7. XML2RDF plus XSD2OWL
• Enrich RDF with links to classes for
corresponding XSD complexTypes:

8. Comparability
• Generate semantic links across accounting
standards (XBRL taxonomies)
– Automatic ontology alignment tools
– Experts input
– Numeric analysis
• Comparability experiment:
– Telefonica S.A. Balance Sheet 2008 and 2009
• XBRL for Spanish CNMV
• XBRL for US SEC

9. Semantic Links US SEC - Spanish CNMV

11. DESCRIBE ?r
WHERE {
?r a ifrs:NoncurrentAssets }
DESCRIBE ?r
WHERE {
?r a
ifrs:OtherNoncurrentFinancialAssets }

12. Conclusions
• XBRL tools: for financial information edition and
maintenance (formulae,…)
• Semantic Web tools: data integration, cross
querying,… COMPARABILITY
– Across filings, companies, accounting principles,…
• However, lack of impact of Semantic XBRL
– Project discontinued 2013
– Too early?
– Too much semantics… or too little?
XSD2OWLXML2RDF

13. Financial Industry Business Ontology
• FIBO™ open industry standard for defining financial
contracts terms, facts and relationships
• 11 core domains, 49 modules and 418 ontology files:
– Foundations
– Business Entities
– Indices and Indicators
– Securities Reference Terms
– Derivatives
– Loans
– Market Data
– Collective Investment Vehicles
– Corporate Actions
– Payments
– Portfolios and Holdings
– …

14. Derivatives / Swaps

15. Experience with US Banks
• Consulting for Cambridge Semantics
– http://www.cambridgesemantics.com
• Banks interest in heavyweight semantics, data
integration for regulators reports automation
• Proof of Concept with Interest Rate Swaps
– Modelling
– Exploration
– Automatic classification
(reasoning)
• Cross-currency swaps
• Single-currency swaps
• Float-Fixed swaps
• …

16. Experience with US Banks

17. Thank you for your attention
Questions?
Roberto García
@rogargon
roberto.garcia@udl.cat
http://rhizomik.net/~roberto/
HCI and Data Integration Research Group
Universitat de Lleida, Spain