Integrating Plant Protection Product Data using Semantic ... · SEMIC 2013 June 21, Dublin . Who we...
Transcript of Integrating Plant Protection Product Data using Semantic ... · SEMIC 2013 June 21, Dublin . Who we...
Integrating Plant Protection
Product Data using Semantic
Technologies
Giorgos Georgiannakis
SEMIC 2013
June 21, Dublin
Who we are
Why Semantic Technologies
• Data rich organisation
• Facilitate sharing of data
• Nature of data
• Openness
• Formats
• Exchange
• Applications
Objective
• Make up-to-date public data available
• Apply standard open data formats
• Share access to the data (humans and machines)
• Provide relevant links to the data
• Facilitate relevant linking to the data
• Develop applications based on the data
The case for plant protection products in EU
A simple yet difficult
to answer question:
“Which products
are authorised for
a crop / pest?’’
DATA CONSUMER
?
Lack of a single point of access
GR
DG Plant Produce
HU
Agrinex
NL
CTGB
PL
BIP
SE
KEMI
AT
AGES
BE
Fytoweb
DE
BVL
Data fragmentation Heterogeneous data formats
Lack of common identifiers
ISA Activity
Domain ontology
(DG Health and
Consumers) Data Cleansing
Common
Taxonomies
Member States
Data / Systems
Common Data Models RDF Repository
SPARQL Faceted browsing
Services
DG Health and Consumers Activities
• Open Data
• Linked Data
• Architecture and infrastructure
• Linking and publishing data from external sources
• MS Authorities
• International authorities
Make up-to-date pubic data available
• Legal obligation
• Data inventory (lack of interoperability)
• Heterogeneous data
• Data exchange and sharing national datasets
• Open Data Portal http://open-data.europa.eu/
Setting an Architecture with Standard Open Data Formats
UI – publication
SPARQL User Interface
Linked Data Tools – Data Dictionary Maker
• Web app
• Create information models
• Describe content of datasets
• Use restricted value lists
• Publish this model on the web for data collection
Linked Data Tools - RDFa Maker
• Transform a CSV dataset into a mini-RDFa, static
website
• Generates an index (URI) to the dataset
• Publish on a local webserver
• Each row of the CSV file must contain:
• Id (unique and permanent identifier)
• Label
• Description
• Datasets are available as human or machine-readable
Linked Data Hub
The process for publication and linkage
• Original data
• Modeling concepts
• Transformation
• Linkage
• Publication
• Applications
• Content Maintenance and evolution
Messages
• Difficulties to link data from multiple sources
• Value of
• Ontologies
• Taxonomies
• Vocabularies
• Technologies are available and mature
• Build for evolution
• Model
• Data
• Practical solutions vs. huge projects
Thank you!