Stats on the Web Best Practices subgroup call

24 January 2018

Meeting Minutes

Use cases

<billroberts> https://‌github.com/‌w3c/‌sdw/‌blob/‌gh-pages/‌stats-bp/‌draft-use-case-list.md

<billroberts> https://‌lists.w3.org/‌Archives/‌Public/‌public-sdwig/‌2018Jan/‌0048.html

ChrisLittle_: +1 pointing out SDMX has been around with international statistical best practice - foundation for RDF-QB - other stuff never made it - need to understand scope and limitations

billroberts: notes SDMX is difficult to read 250 page type doc...

AndreaPerego: analysis was a quick and dirty job, statistical data not core business...
… as a template to identify where we have BP already, where gaps
… URIs not part of of SDMX - have internally scoped identifiers
… need to provide guidance here
… discussion about not being too RDF biased - how do we apply - e.g., what about CSV on Web? Can it play a role for making stat data webby?
… BP work identifies number of possible options
… some BP were about metadata and some about data
… RDF-QB can be about both
… SDMX metadata is often free text - although often it describes some information very accurately, and covering different aspects (e.g., data quality)

<billroberts> (Yes I think that's accurate Rob)

AndreaPerego: should at least address DWBP cllauses

toba: SDMX has a lot of different things mixed in one standard, which is what makes it complex. RDF-QB pulled out the core model of it, without doing all the codelists
... which I think was the right thing to do. But leaves a gap in that the many codelists in SDMX are not a core part of RDF-QB , but they are not necessarily stably governed
... so there is a general challenge for these kind of BPs of how do we put up standard maintained vocabs that can be re-used
... When looking at QB4ST, i deliberately scoped it to metadata (not data) and considered it as a canonical version of a conceptual model. Could certainly have been in other syntaxes
...JSON-LD version of RDF-QB is quite straightforward
.... We don't have to solve all problems: a reasonable scope would be metadata to describe whatever representation we have, using RDF-QB as core, and allowing RDF or JSON expression of it

<ChrisLittle_> +1 to focus on metadata

billroberts: agree we shoudl focus on web specific aspect

ChrisLittle_: notes QB4ST perspective that the "big data cubes" arenot going to leave their existing implementation choices for data encoding and access mechanisms

billroberts: DXWG needs folllowing

<billroberts> http://‌ggim.un.org/‌meetings/‌GGIM-committee/‌documents/‌GGIM6/‌Background-Paper-Proposal-for-a-global-statistical-geospatial-framework.pdf

<AndreaPerego> Just to say UNGGIM was also mentioned during the SDW BP call.

roba:Data Exchange WG has just released first draft of requirements. Not had much feedback yet
...would be great to have some feedback from this group to DXWG (and not from Rob as he is an editor of it)
...so please everyone do send comments to DXWG on that requirements doc
...Description of profiles and negotiation of which one you get is something DXWG is looking at
...It's more than negotiating data format, but rather negotiating by profile - could be a useful angle for us
...DXWG won't define specific profiles - that would be down to other groups, eg SDWIG
...so we could try to come up with a profile for metadata for statistics in that framework

<AndreaPerego> +1

https://‌www.w3.org/‌TR/‌dcat-ucr/ has details on how to comment in intro

josephabhayaratna: has a set of comments from ABS on this groups UC, yet to write up and submit

marqh: helpful to have explicit examples

billroberts: using RDF-QB in real world - can bring forward some examples

marqh: sounds really helpful

Action: bill roberts bring forwards RDF-QB examples from experience

<trackbot> Created ACTION-380 - Roberts bring forwards rdf-qb examples from experience [on Bill Roberts - due 2018-01-31].

F2F meeting in Amersfoort

<billroberts> https://‌w3c.github.io/‌sdw/‌meetings/‌f2f-1.html

-1 for F2F

<ChrisLittle_> Chris planning to

<AndreaPerego> Not sure I can make it... But try to join remotely.

<josephabhayaratna> I won't be there

roba: have some kind of plenary catchup that they could dial in to but not very productive to join a F2F meeting remotely

roba: if possible, spend some time reviewing possible examples and identify the most useful to get principles across to audience

<ChrisLittle_> +1 to >1/2 day

roba: I've been working as an OGC staff member to bring the definition publication capability in-house from a skunkworks project that SImon Cox started
... so there should be a crude CMS able to publish and maintain registers of terms as linked data. Hence OGC will be able to publish vocabs on behalf of working groups and something we can take advantage of
...The process would be to approach the naming authority to ask for a namespace. Early days but should be possible
...eg for SDMX codelists say
...ISO TC211 might use this
...can include automatic entailment, eg of broader/narrower relations
...but main decisions are around governance and that will be done on a case to case basis

<ChrisLittle_> bye

<AndreaPerego> Bye!

<billroberts> bye all, thanks

Summary of Action Items

  1. bill roberts bring forwards RDF-QB examples from experience
Minutes formatted by Bert Bos's scribe.perl version 2.37 (2017/11/06 19:13:35), a reimplementation of David Booth's scribe.perl. See CVS log.