Semantic Govenment Vocabulary Petr Křemen Project: Implementace strategií v oblasti otevřených dat II Reg. number: CZ.03.4.74/0.0/0.0/15_025/0004172
Problem 1 Keyword ambiguity 2 Buildings
Problem 1 Keyword ambiguity 3 Building is a construction ● above ground ● with solid foundations ● spatially compact ● with walls and roof Act 406/2000 Coll., on Energy Management Act 256/2013 Coll, Cadastral Law Building is a construction ● both above and below ground ● spatially compact ● with walls and roof ● with heating Bus stops ?
Problem 2 Lack of semantic relationships between keywords ● How to find semantically interelated datasets ? Political Parties in the Chamber of Deputies Registry of candidate lists for Parliamentary elections 2017 Registry of candidates for Parliamentary elections 2017 Banking accounts of political parties Donations for elections to the Chamber of Deputies Members of the Chamber of Deputies Votes in the Chamber of Deputies Votes of members of the Chamber of Deputies
Problem 3 Discovering related datasets Building Building (process) Building (object) Building (256/2013) Foundations has part Object Event has participant Building (406/2000)
Semantic Government Vocabulary (SGoV) 6
SGoV Architecture 7 ● Basic Vocabulary ○ E.g. Event, Kind, Relator ○ Based on the Unified Foundational Ontology ● Vocabulary of Public Sector ○ E.g. Legal Subject, Organization ○ Mapped to ISA Core Vocabularies ● Vocabularies of Legislation ○ E.g. Building (Act 406/2000 Coll., on Energy Management) ○ 1 vocabulary per Act ● Vocabularies of Agenda ○ 1 vocabulary per Agenda ● Vocabularies of Datasets ○ 1 vocabulary per Data Schema (to support data series)
Problem 3 Discovering related datasets Building Building (process) Building (object) Building (256/2013) Foundations has part Object Event has participant Building (406/2000)
Problem 3 Discovering related datasets Legal Object Building (process) Building (object) Building (256/2013) Foundations has part Object Event has participant Building (406/2000) Vocabulary of Legislation Vocabulary of Agenda Vocabulary of Public Sector Basic Vocabulary
Unified Foundational Ontology 10
Vocabulary of Public Sector 11 ● Modeled using OntoUML ○ Inspired by OntoClean ● Aimed at describing common-sense notions in the public Sector ● To be reused by multiple vocabularies
Leg./ Agenda Vocabularies 12 ● Modeled using the Vocabulary of Public Sector ● Aimed at describing legal/agenda terms necessary to describe the meaning of a dataset
SGoV Vocabulary Internals 13 ● Each vocabulary is in OWL2-DL (expressiveness ALCNI) ○ Glossary ■ SKOS ○ Ontological model ■ OWL2-DL
Search in NKOD Proposal 14 ● Dataset vocabularies are interconnected through the stack of legal and agenda vocabularies ● Different metrics to find out similar datasets. Two datasets are relevant if ○ they share same terms ○ they share same object types/event types ○ etc. ●
- inference
Current Vocabulary Usage 15 ● Open data of the Register of rights and responsibilities of public authorities ● Experiment on the domain of Parliamentary Elections in the Czech Republic ○ 3 people - 4 hours long training for them ■ Using only basic language (not OntoUML) ● Object Type ● Event Type ● Relator Type ● Intrinsic Trope Type ● + links according to UFO ○ Partial vocabularies for 40 Acts ○ Over 1000 terms
SGoV Resources 16 ● https://slovník.gov.cz/sparql ○ SPARQL endpoint over SGoV ● https://slovník.gov.cz/prohlížeč ○ simple faceted browser of SGoV ● https://github.com/opendata-mvcr/ssp
Summary 17 ● Main ideas ○ Layered architecture ■ To have the vocabulary lifecycle managable ○ Some well-founded top-level/public sector ontology ■ To define metrics for dataset similarity ○ Legislation-founded vocabularies ■ To be used for (not only open) data description ○ Broader use ■ Share meaning of terms used in eGovernment ■ Share meaning of datasets (rich metadata)