Spatial Planning On The Semantic Web Terracognita 2009

Spatial Planning on the Semantic Web

Radboud Winkels, Rinke Hoekstra, Erik Hupkes

04/10/2023 2

Outline

• Use Case• Requirements– Metadata– Legal perspective

• Representation– Example

• Limitations• Future work

04/10/2023 3

04/10/2023 4

Room for Improvement

• Finegrained coupling with regulation texts

• Assessment– “Am I allowed to build X here?”

• Planning– “Where am I allowed to build X?”

• Evaluation– “Are land use regulations X and Y consistent?”– “What is the impact of policy X?”

Legal Perspective

• Plans do not describe existing situation, but • Prescribe restrictions and rights associated with

geospatial objects.

• Maps provide intuitive handles for evaluating the normative content of land use regulations:– Hierarchical relation between authorities coincides with

spatial inclusion– Adjacency determines indirect effect (e.g. industry next

to nature reserve)04/10/2023 5

Heterogeneity

• Spatial plans: maps and documents• Maps & other data– Geotagging of photos, traffic conditions &c.

• How to ‘link’ the data?– Determine meaningful overlap between data from

involved domains– Spatial plans: • Land use category

04/10/2023 6

Metadata and Ownership• Multiple origins– Regulations issued by different government bodies with

overlapping jurisdiction– Different categorisation schemes– Subject to change

• Problems– Overlap: users unaware of interaction between

jurisdictions• Transparent presentation of all applicable regulations on a single

map

– Comparison: different schemes hinder comparison of land use regimes• Promote sharing of categorisation schemes

04/10/2023 7

Approaches to scheme sharing

• Standardisation • Top-down, prescribed, single domain

• IMRO 2006: obligatory for municipal urban planning in NL• INSPIRE:

standardise exchange of spatial information within EU

• Umbrella, unifying framework, multiple domains

• GEMET: multilingual thesaurus of 5000+ environmental terms

04/10/2023 8

Requirements (1)

• Categorisation scheme maintained by owner of standard

• Integration requires a mapping between categorisation schemes– Flexible mechanism (cf. BestMap paper at OWLED)

• Representation maintained by owner of regulation

• Distributed content and semantics– Semantic Web

04/10/2023 9

Requirements (2)

• Maps– Served from standards compliant web services– Accessible through off-the-shelf API’s

• Regulation texts– Served from web-accessible locations, – In a format that allows for integration with metadata

• Metadata– Used both for information on maps and corresponding texts– Mapping between vocabularies

• Norms– Represented in terms of standard metadata, – Expressed using a Semantic Web compliant language

04/10/2023 10

LegalAtlas (a.o. Winkels et al., 2007)

Connection between texts and map Represented in RDF

Not web based Not service based Limited to IMRO 2006 categories No representation of norms No normative reasoning Annotation mechanism:

`Designations’ are annotated with maps and texts, instead of the other way around

Not extensible with other land use categorization schemes Land use categories as classes: lots of individuals

04/10/2023 11

Information Serving

• Geographic Information Systems– Exchange• GeoRSS, GML, KML, ESRI Shapefiles

– Services• Web Map Service (WMS), Web Feature Service (WFS)• Basic geospatial reasoning facilities

• No support for normative aspects of spatial plans

04/10/2023 12

Towards a Legal Information Server

• Provide normative reasoning as a service• Assess whether some situation is allowed or

disallowed, given a set of norms

• Implementations:– LIS (Winkels et al., ‘98,’02)

• SWI-Prolog

– LKIF Core (Hoekstra et al. ’09, Hoekstra ’09) andHARNESS (van de Ven, et al. ‘08)

• OWL 2 DL

04/10/2023 13

• Revised annotation mechanism– Region-centred– Land use categories as SKOS concepts– Mappings between categories as OWL classes (BestMap, OWLED)

• Web service based– Sesame RDF triple store– SwiftOWLIM Reasoner– OpenGIS GeoServer– Google Maps

• Spatial norms– Allowed/Disallowed situations– OWL 2 DL class descriptions

Prototype System

04/10/2023 14

04/10/2023 15

Flevoland

04/10/2023 22

Using OWL 2 DL…• Advantages– Standard, ‘off the shelf’ reasoners– Decidable, complete, guaranteed response– Monotonic

• Disadvantages– Limited expressiveness vs. complexity of world• OWL 2 DL is restricted to ‘tree models’• Complex configurations of objects are hard to define

– No obvious way to connect DL reasoning to GIS

04/10/2023 23

Conclusions

• Representation of distributed spatial regulations

• Map based legal assessment as web service• Compare and evaluate spatial regulations

• Application of other work to new domain– Legal case assessment method in OWL 2 DL– Ontology mapping (BestMap, OWLED)

04/10/2023 24

Future Work

• Explaining Results • Query for conflicts– Currently only for specific cases

• More advanced reasoning – “Space Package” (Hage et al. 2009)– PelletSpatial (RCC8, Stocker & Sirin, 2009)– Dealing with exception hierarchies

(GIS Transactions)

04/10/2023 25