Warning:
This wiki has been archived and is now read-only.

Linked-data

From Spatial Data on the Web Working Group
Jump to: navigation, search

Where do links commonly occur (in spatial data)?

  • Statistics data associated to location
    • data associated to latitude
    • data associated to longitude
  • Data or information related to complex spatial data (e.g. a map of school locations should be linked to location of all schools used)
  • Other spatial data on the same location or area available in other scales
  • Thematic data/concepts related to the spatial data
  • Provenance data including source, device, platform related information
  • Temporal data (time/zone/sequence)
  • Ownership information
    • Copyright and license
    • Access right data
  • Administrative/cultural relationships between places e.g capitals
  • Source data geometry link to interpolated/interpreted line/area/volume
  • Connectivity relationships – network distance, connectedness, line of sight
  • Dependency relationship where the geometry of one object defines part of another e.g. hydrographic features used to define administrative boundaries
  • Topological relations in 3rd dimension – e.g. pipeline is below a road, flightpath is over a city. i.e. geometries are disjoint but they overlap in the horizontal plane.

What should I link to? (or link between)

  • Link between datasets
    • and also link to/between subsets and (sometimes) individual data items (e.g. link to slices of a coverage)
    • bridge between identifiers and access to actual data (subsets). (e.g. creating an RDF fragment that describes the subset, and coupling with a WCS URL that accesses it)
  • Link to *features* that we extract from the data
  • Link to representations in other formats (eg, JPEG delivery for an image stored originally as TIFF or NetCDF)
  • Link to an authoritative database of boundaries (geometries) representing accepted definitions of countries, regions, continents, oceans, seas and so forth.

What do the links mean?

  • in the spatial domain links represent spatial relationships of hierarchy (e.g. x is constituent part of y) and topology (e.g. y is adjacent z)
  • We can use "spatial predicates" ([link: https://en.wikipedia.org/wiki/DE-9IM#Spatial_predicates]) to describe spatial links
    • [Krzysztof] Combining spatial and temporal relations into a common system would lead to a huge amount of relations and most of them would not be distinguished by humans (which was already shown for multiple spatial relations).
  • To make use of links, you need to understand what those relationships mean and define/identify the common vocabularies that should be used to describe links
    • Are the link types particular to a given domain?
    • Is it possible to specify general purpose definitions?
    • Can domain-specific vocabularies be mapped to common vocabularies (& if so, how)?
  • use unique, global, durable identifiers
  • links should be typed (implicitly or explicitly) so that client applications can decide which links to follow when traversing a web of interlinked resources to reach application goals
  • issues:
    • need to be able to discover resources so that you can link to them
    • how to assert and maintain links between large sets of resources?
    • how are links expressed? (in a doc, in a link header)
    • how are links characterised? (e.g. typing)
    • how are hints about the target resource conveyed?

How can I describe links in my format of choice?

  • This is the case where we are linking the geometry of a thing (e.g. the polygon of Loch Lomond) to the thing or two sets of geometries of the same thing.
  • We want people creating spatial data to continue to use their existing tool chains.
  • It should be defined in an appropriate level of abstraction
  • RDF (& concrete encodings: TTL, JSON-LD etc.)
  • JSON ... and variants such as GeoJSON
  • XML ... and variants such as GML, KML
  • HTML ... using structured mark-up (RDFa, microdata or embedded JSON-LD)
  • CSV (building on the CSV on the Web now at CR [1]


[1]: Tabular Data Model [1], Tabular Metadata [2], CSV2RDF [3] and CSV2JSON [4]

What else links to my subject of interest?

  • Statistical data (expressed in RDF Data Cubes for example) often relates to geographic and administrative areas.

Many statistical datasets may use the same "spatial" dataset.

  • A common use case is to look for other statistics relating the subject administrative area.
  • Authoritative reference data.

For example, governmental data sets that provide elaborate, high quality data on administrative areas (including geometry and its history of course), which can be reused by anyone wanting to say or publish something about such an area. With such authoritative reference data in place it would be easy to use the same identifiers for locations, and to find data about those locations on the web.

How can I manage links between sets of things (so that I don't have to manage individual links)?

  • We can use VoID (and Linksets) to say that "this" dataset is related to "that" dataset,

but then we still need create the links _between_ the individual resources that those datasets describe so we can browse / traverse between related resources.

  • Using a set of rules and categories of properties (e.g. find all the linked data sets that provide environmental monitoring data for an area)

How can third parties assert links and relationships (and should I trust their assertions?)

  • Related questions:
    • How can metadata be associated with a link statement ? (who made the statement, when the statement was made, how certain it is, temporal extent of validity).
    • How can I assert something about a third party’s link statement e.g. if I want to rate it as validated/wrong/dubious ?
  • An oa:Annotation (as in W3C Open Annotation) is a reified link, to which we can attach further metadata (who made the annotation, why, when, etc).
Retrieved from "https://www.w3.org/2015/spatial/wiki/index.php?title=Linked-data&oldid=1296"