The National Map as Knowledge Base Topographic Linked

The National Map as Knowledge Base Topographic Linked Open Data Dalia Varanka Research Physical Scientist Center of Excellence in Geospatial Information Science With: Matthew Wagner, Tanner Fry, Jacques Bourquin, and Lynn Usery January 8, 2020

The National Map as Knowledge Base Concept The idea behind the prototype Objective: Enable topologic & browseable graph approaches on internal and external linked data using interfaces for maps, tables, concept diagrams, and other media to build relevant knowledge subgraphs Design Stages • Instances populated through services • Ontology integration (co-references) • User-centered design interfaces

Focus on Data Summaries All URIs are dereferenceable; functional subsets are aggregated Best Practice • Use data associated by multiple property links • Use data associated through transitivity, to provide description • Avoid associated descriptions forming redundant or implication relations • Use only one of alternative URIs to associate different values for the same property

Trends in Linked Open Data (LOD) Implications of co-referencing data by owl: same. As Problem • Entities linked by owl: same. As may refer to the same thing, but their graphs may conflate conflicting context-dependent descriptions through inference • Potentially conflicting literal values for same or similar properties, e. g. , five vs nine digit zip codes • owl: same. As refers to ‘identity’ and statements should be transitive, symmetric, and reflexive.

Linked Open Data – Simple Ontologies Graphs describing feature instances Modules for different themes Ontology Pattern • Gazetteer-type network • Geo. SPARQL ontology • Complex feature cross-thematic linking

Place-Based Semantics Multi-thematic set integration US Topo Background and LD source Data • 7. 5 -minute quadrangles • Layered Geo. PDF • Orthoimage base • Core feature layers Ortho-rectified Aerial Imagery Roads Names Elevation Contours Hydrography State/County/USFS Boundaries Runways Woodland Railroads PLSS Fire Stations Hospitals Schools Military Boundaries Cemeteries Post Offices Shaded Relief Selected Trails FWS Boundaries State Capitals Police Stations Correctional Facilities IMBA Trails National Scenic Trails

Advanced Feature Description Integrates The National Map with data from diverse sources Advances • Accepts multiple data formats • Draws on services for input • Dereferences data

Feature Types/Feature Codes Feature type look-up tables Geospatial features are geometry objects in GIS Knowledge graphs require ontologically specific feature types Inference model integrates • Coordinate geometries • Feature codes • Ontology classes and properties

Feature Type Ontology / GIS Geometry Layer Integration Hydrographical feature gazetteer Geospatial geometry layers Cognitive ontological features Inference linking • Feature types • Geographic coordinate geometry • Lookup table codes

LOD System for Topographic Data Integration Summary • Knowledge base design from relational table sources • Cross-thematic links for complex geospatial features • Inputs from other USGS products and external sources • User interface for cognitive and automated reasoning

The National Map as Knowledge Base Topographic Linked Open Data Dalia Varanka dvaranka@usgs. gov https: //www. usgs. gov/core-science-systems/ngp/cegis/geospatial-semantics-andontology January 8, 2020