sdmx-attribute:coverageTime

[Leigh Dodds]

Hi,

Is anyone using sdmx-attribute:coverageTime? The SDMX definition says
its free text so we might expect something like "2012-2000" but I
wondering if a time interval might be better

@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix tl: <http://purl.org/NET/c4dm/timeline.owl#>.
@prefix qb: <http://purl.org/linked-data/cube#> .
@prefix sdmx-attribute: <http://purl.org/linked-data/sdmx/2009/attribute#> .

:eg a qb:Dataset;
rdfs:label "Dataset with start/end dates for coverage";
sdmx:attribute:coverageTime :interval.

:interval
a tl:Interval;
tl:start "2000"^^xsd:Year;
tl:end "2012"^^xsd:Year.

Any thoughts?

L.

--
Leigh Dodds
Freelance Technologist
Open Data, Linked Data Geek
t: @ldodds
w: ldodds.com
e: le...@ldodds.com

[Dave Reynolds]

On 10/10/12 14:18, Leigh Dodds wrote:
> Hi,
>
> Is anyone using sdmx-attribute:coverageTime? The SDMX definition says
> its free text so we might expect something like "2012-2000" but I
> wondering if a time interval might be better
>
> @prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
> @prefix tl: <http://purl.org/NET/c4dm/timeline.owl#>.
> @prefix qb: <http://purl.org/linked-data/cube#> .
> @prefix sdmx-attribute: <http://purl.org/linked-data/sdmx/2009/attribute#> .
>
> :eg a qb:Dataset;
> rdfs:label "Dataset with start/end dates for coverage";
> sdmx:attribute:coverageTime :interval.
>
> :interval
> a tl:Interval;
> tl:start "2000"^^xsd:Year;
> tl:end "2012"^^xsd:Year.
>
> Any thoughts?

Agreed that a time interval would be better. There's a few places where
SDMX allows free text but where in Linked Data cases we'd tend to have
structured vocab.

There is also dct:temporal, which does expect a resource, as an alternative.

For the interval itself I might be inclined to use interval:Interval
(Stuart's ontology for the reference time service [1]) which in turn is
a subclass of owltime:DateTimeInterval. Though tl: does look simpler.

Dave

[1] http://reference.data.gov.uk/def/intervals/

[Sarven Capadisli]

+1 to intervals like [1].

In fact, I think it works out quite well, even for different types of
intervals within the same observation. Example:

http://worldbank.270a.info/dataset/world-bank-climates/month-average-anomaly/2060-2079/08/3/tas/ipsl-cm4/b1

has an interval that has a coverage (20 year period starting from year
2060) and another interval that states the actual occurring interval
(month of August).

-Sarven