![Federated Querying using TPF
Data sources
● http://linked.open.gent/parking/
● http://data.linkeddatafragments.org/linkedgeodata
Query (link)
SELECT ?name ?free ?wktString WHERE {
?ParkingGent a <http://vocab.datex.org/terms#UrbanParkingSite>.
?ParkingGent owl:sameAs ?parkingLGD .
?ParkingGent rdfs:label ?name.
?parkingLGD <http://geovocab.org/geometry#geometry> [
<http://www.opengis.net/rdf#asWKT> ?wktString] .
?ParkingGent <http://vocab.datex.org/terms#parkingNumberOfVacantSpaces> ?free .
FILTER ( ?free > 200 )
}](https://image.slidesharecdn.com/otn-demolinkeddatareal-timeparkinggent-170125170243/85/Real-time-Linked-Transport-Data-in-Ghent-9-320.jpg)
Real-time Linked Transport Data in Ghent
- 1. Real-time Linked Transport Data in Ghent Dieter De Paepe Thimo Thoeye Pieter Colpaert
- 2. A virtual Traffic Centre in Ghent
- 5. Open Data Challenges Semantic domain modelling Making sure we’re talking about the same things Scaling Providing a stable data source for many users Queryability Allowing users to ask broad questions
- 6. Example: Parking spaces Semantic domain modelling There are 4 departments that manage parking spaces that are “accessible for people with disabilities”. Social service, roadworks, etc. Each uses a different way of storing this information and terms are not re-used. Scaling Slowly, we are connecting more and more data sources to the triple store, which forces services to think about shared semantics and concepts. Nevertheless, a triple store is not scalable. We need to find a way in which semantics can be fed back into the existing processes and tools (GIS, etc.) Queryability Right now, querying is done on the scale of 1 service/dept. Federated querying remains a huge exercise and should definitely be improved.
- 7. Solution: a large-scale information system Linked Data use URIs for identifying and documenting things HTTP with REST constraints ● Uniform interface ● Stateless ● Cacheable Hypermedia based querying
- 8. Use case: availability of parking sites in Ghent 1. Determine suitable RDF ontologies 2. Map live data to these ontologies 3. Add metadata and hypermedia 4. Set up REST enabled webpage
- 9. Federated Querying using TPF Data sources ● http://linked.open.gent/parking/ ● http://data.linkeddatafragments.org/linkedgeodata Query (link) SELECT ?name ?free ?wktString WHERE { ?ParkingGent a <http://vocab.datex.org/terms#UrbanParkingSite>. ?ParkingGent owl:sameAs ?parkingLGD . ?ParkingGent rdfs:label ?name. ?parkingLGD <http://geovocab.org/geometry#geometry> [ <http://www.opengis.net/rdf#asWKT> ?wktString] . ?ParkingGent <http://vocab.datex.org/terms#parkingNumberOfVacantSpaces> ?free . FILTER ( ?free > 200 ) }
- 10. And this can be used in a user-interface plotting the updating availability based on this federated query on a map Tweakable code example: http://codepen.io/pietercolpaert/pen/wgoPNV