There have been some discussions about reuse of statistics with people from Statistics Norway.[1] They use a format called JSON-stat.[2] A bunch of census bureaus are starting to use JSON-stat, for example Statistics Norway, UK’s Office for National Statistics, Statistics Sweden, Statistics Denmark, Instituto Galego de Estatística, and Central Statistics Office of Ireland. I've heard about other too.
I have started on some rant at Meta about it, I didn't finish it.[3] Perhaps more people will join in? ;)
A central problem is that statistics are often produced as a multidimensional dataset, where our topics are only single indices on one of the dimensions. We can extract the relevant data, but it is probably better to make a kind of composite key into the dataset to identify the relevant stuff about our topic. That key can be stored as a table-specific statement in Wikidata, and with a little bit of planning it can be statistics-specific or even bureau-specific.
[1] https://ssb.no/en/ [2] http://json-stat.org/ [3] https://meta.wikimedia.org/wiki/Grants:IdeaLab/Import_and_visualize_census_d...
On Fri, Aug 7, 2015 at 3:06 PM, Joe Filceolaire filceolaire@gmail.com wrote:
Hypercubes and csv flat files belong in commons in my opinion (commons may have a different opinion ). That's if we even want to store a copy.
This source data should then be translated into wikidata triples and statements and imported into wikidata items.
The statements in wikidata are then used to generate lists and tables and graphs and info graphics in wikipedia.
At least that's how I see it
Joe
On Thu, 6 Aug 2015 17:00 Jane Darnell jane023@gmail.com wrote:
I have used the CBS website to compile my own statistics for research. Their data is completely available online as far as I know and you can download the queries you run on the fly in .csv file format, or text or excel. They have various data tables depending on what you find interesting and complete tables of historical data is also available. That said, I think any pilot project would need to start with their publications, which are also available online. These can be freely used as sources for statements. Interesting data for Wikidata could be population statistics of major cities per century or employment statistics per city per century and so forth. See CBS.nl
On Thu, Aug 6, 2015 at 5:30 PM, Gerard Meijssen gerard.meijssen@gmail.com wrote:
Hoi, As far as I am concerned, data that is available on the web is fine if you use data on the web. It makes no difference when the data is to be used in the context of the WMF.
When the CBS shares data with us in Wikidata, it makes the data available in Wikipedia.
It is why I would like for something small, a pilot project something where we can build on. Thanks, GerardM
On 6 August 2015 at 17:23, Thad Guidry thadguidry@gmail.com wrote:
Netherlands Statistics should just post the data on the web...so that anyone can use its "Linked Data".
And actually, CSV on the Web is now a reality (no longer a need for XBRL)
https://lists.w3.org/Archives/Public/public-vocabs/2015Jul/0016.html
As DanBri notes in his P.S. at the bottom of the above link..."... the ability (in the csv2rdf doc) to map from
rows in a table via templates into RDF triples....is very powerful."
Thad +ThadGuidry
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata