Exploratory querying of the Dutch GeoRegisters
•Download as PPTX, PDF•
0 likes•201 views
This document introduces exploratory querying and SPEX, a tool for exploratory querying of spatial and temporal data. It summarizes the goals of exploratory querying, gives examples of exploratory querying software, and demonstrates SPEX through use cases of exploring integrated datasets like the Dutch BAG registry and real estate listings. The presentation describes SPEX's capabilities for interactively querying and visualizing linked geospatial and temporal data to understand dataset contents for further use and integration.
More Related Content
What's hot
What's hot (20)
Similar to Exploratory querying of the Dutch GeoRegisters
Similar to Exploratory querying of the Dutch GeoRegisters (20)
Exploratory querying of the Dutch GeoRegisters
- 1. Exploratory querying of the Dutch Georegisters with the purpose of further integration with other sources by Stanislav Ronzhin & Rob Lemmens
- 2. • To introduce Exploratory Querying • To present SPEX, a tool for Exploratory Querying in space and time • To demonstrate SPEX in action Goals for today
- 3. About myself.. • 2002-2007 • 2008 - 2013 • 2013-2015 http://frecom.ru/ Stanislav Ronzhin
- 4. The pathetic fallacy of RDF
- 5. Racefietsen example • Frame material • № of gears • Brakes • ……
- 7. Exploratory Querying software Faceted browsers: • LESS • RelFinder • gFacet • RDF Gravity • Tabulator • Rhizomer Visual SPARQL clients: • LodLive • NITELIGHT • IsaViz (SPARQLViz) • ViziQuer • OpenLink iSPARQL • Sgvizler • QueryVOWL • Visualbox • SparqlFilterFlow
- 8. Exploratory Querying of spatio-temporal data
- 9. Exploratory Querying of spatio-temporal data bagGeopunt:0546010000010756 geo:asWKT "POINT(4.4884203637992 52.157846104773)". bagWoonplaats:2088 a bag:Woonplaats; bag:woonplaatsnaam "Leiden“. bagPand:0546100000040803 a bag:Pand; bag:bouwjaar “1650”; bag:ingangsdatum "2010-08-26T00:00:00".
- 10. SPEX - Spatio-temporal Content Explorer Scheider, S., Degbelo, A., Lemmens, R., van Elzakker, C., Zimmerhof, P., Kostic, N., Jones, J.,& Banhatti, G. (2015, in publishing). Exploratory querying of SPARQL endpoints in space and time. Semantic Web journal.
- 11. Who are the users? • Data managers, (geo) information professionals, non-experts (with a little help) • Who want to understand the content of data for further use/integration • Unexperienced semantic web/SPARQL users
- 12. Emergency management use case • Browser for triplified and enriched Ushahidi data • Selection of operating hospitals in some area
- 13. Use case: BAG + HuizenZoeker + Energie labels
- 14. Use case demonstration 1. Exploration of BAG 2. Exploration of other datasets for the sake of further integration
- 15. Use case: BAG + HuizenZoeker + Energie labels
- 16. Query to select all the verblijfsobjects with their area in a neighborhood of interest
- 17. Query to select all the verblijfsobjects with their addresses in a neighborhood
- 18. To sum up • Exploratory Querying – simultaneously learning about the information needed while specifying it • SPEX is a prototype tool for Exploratory Querying in space and time • For those who want to know the content of data for further use/integration
- 19. Future development • Develop workflow that would embed SPEX • Named Graph support, k-Nearest Neighbor query, functionality for data extraction
Editor's Notes
- Why I changed the title?
- great big graphs have severe drawbacks for visualizing linked data sets because what they exactly not provide are overview, zoom and details-on-demand.
- Think about a novice who is willing to buy his or her first road bicycle. The problem here is that for a novice all of them look almost identical when searching on the WEB. In order to find a suitable one a novice has to deal with an old dilemma of joining the learning of concepts about bikes together with their specification. In the case of a bike search, a novice learns about different characteristics of a road bicycle (e.g frame material, number of gears, etc.) while exploring the range of possible variants. When the biker finally decides on the particular value of a characteristic (e.g aluminum frame), he or she is specifying this value.
- In contrast to relational databases, linked data is self-descriptive. This means that linked metadata are just additional triples that are stored together with other data triples. Therefore, it allows doing both specification and learning about concepts at the same time. For instance, linked data representation of BGT data (Basisregistratie Grootschalige Topografie) consists of more than 180 different classes and numerous data triples. The question is how a user can find data of interest, for instance about his or her property. This can be done in a closed iterative loop of classify and instantiate queries (Scheider et al., 2015). The former searches for classes and relations, thus facilitating learning of the BGT concepts when the latter helps to explore particular instances of identified classes and relations. The results of a classify query feeds into an instantiation query, which in turn provide information that feedbacks new classify query and so on. Thus, data exploration involves querying and querying in turn cannot be done without exploration. This approach is called exploratory querying in (Kadlag et al., 2004) and, more generally, exploratory search in (Marchionini. 2006; White & Roth, 2009).
- From these, RelFinder and gFacet directly work on SPARQL endpoints and can be used without a-priori knowledge about content, based on autosuggestion and feedback. RelFinder is restricted to instance based queries without variables. gFacet can already be considered a visual query tool, as it follows a visual graph-pattern strategy very similar to the design principles proposed in this paper (compare Section 4.2). From these, iSPARQL, SPARQLViz, Visualbox and Sgvizler require substantial apriori knowledge about SPARQL or contained vocabularies for building a meaningful query. LodLive, ViziQuer, NITELIGHT, QueryVOWL and SparqlFilterFlow, in contrast, have a form of overview and suggestion tool for available vocabularies and datasets in an endpoint as well as substantial feedback and support in building queries. ViziQuer and NITELIGHT, however, seem to be made primarily for tech-users and focus less on data exploration. An interesting interactive query tool suitable for inexperienced users which gives feedback on data satisfying a query in terms of a “flow” chart is SparqlFilterFlow
- In its most basic diffinition Geoinformation consists of tree components: SPACE TIME and THEME. These roughly resemble What Where When.
- Instead, a mix of graphical options which fit these tasks and corresponding data types are more adequate. Map-like query interfaces can play a key role in retrieval tasks and are easily adopted by users. Visual tool for construction of graph patterns A result set pane, A query pane. Map and a time slider. A visual query system needs to support query formulation by letting users select visual data representation elements and manipulating them. Each possible manipulation needs to translate into a syntactical operation in the formal query language. Even though the choice of a visual query interface depends on the query language (i.e., SPARQL), the development in the Semantic Web illustrates a wide variety of query construction approaches.